Building machines that learn and think like people

Seeing the random forest through the decision trees. Supporting learning health systems from histopathology with machine learning models: Challenges and opportunities

This paper discusses some overlooked challenges faced when working with machine learning models for histopathology and presents a novel opportunity to support "Learning Health Systems" with them. Initially, the authors elaborate on these challenges after separating them according to their mitigation strategies: those that need innovative approaches, time, or future technological capabilities and those that require a conceptual reappraisal from a critical perspective. Then, a novel opportunity to support "Learning Health Systems" by integrating hidden information extracted by ML models from digitalized histopathology slides with other healthcare big data is presented.

The role of action concepts in physical reasoning: insights from late childhood

A fundamental component of human cognition is the ability to intuitively reason about behaviours of objects and systems in the physical world without resorting to explicit scientific knowledge. This skill was traditionally considered a symbolic process. However, in the last decades, there has been a shift towards ideas of embodiment, suggesting that accessing physical knowledge and predicting physical outcomes is grounded in bodily interactions with the environment. Infants and children, who learn mainly through their embodied experiences, serve as a model to probe the link between reasoning and physical concepts. Here, we tested school-aged children (5- to 15-year-olds) in online reasoning games that involve different physical action concepts such as supporting, launching and clearing. We assessed changes in children's performance and strategies over development and their relationships with the different action concepts. Children reasoned more accurately in problems that involved supporting actions compared to launching or clearing actions. Moreover, when children failed, they were more strategic in subsequent attempts when problems involved support rather than launching or clearing. Children improved with age, but improvements differed across action concepts. Our findings suggest that accessing physical knowledge and predicting physical events are affected by action concepts, and those effects change over development. This article is part of the theme issue 'Minds in movement: embodied cognition in the age of artificial intelligence'.

An algorithmic account for how humans efficiently learn, transfer, and compose hierarchically structured decision policies

Learning structures that effectively abstract decision policies is key to the flexibility of human intelligence. Previous work has shown that humans use hierarchically structured policies to efficiently navigate complex and dynamic environments. However, the computational processes that support the learning and construction of such policies remain insufficiently understood. To address this question, we tested 1026 human participants, who made over 1 million choices combined, in a decision-making task where they could learn, transfer, and recompose multiple sets of hierarchical policies. We propose a novel algorithmic account for the learning processes underlying observed human behavior. We show that humans rely on compressed policies over states in early learning, which gradually unfold into hierarchical representations via meta-learning and Bayesian inference. Our modeling evidence suggests that these hierarchical policies are structured in a temporally backward, rather than forward, fashion. Taken together, these algorithmic architectures characterize how the interplay between reinforcement learning, policy compression, meta-learning, and working memory supports structured decision-making and compositionality in a resource-rational way.

Identifying nucleotide-binding leucine-rich repeat receptor and pathogen effector pairing using transfer-learning and bilinear attention network

MOTIVATION

Nucleotide-binding leucine-rich repeat (NLR) family is a class of immune receptors capable of detecting and defending against pathogen invasion. They have been widely used in crop breeding. Notably, the correspondence between NLRs and effectors (CNE) determines the applicability and effectiveness of NLRs. Unfortunately, CNE data is very scarce. In fact, we've found a substantial 91 291 NLRs confirmed via wet experiments and bioinformatics methods but only 387 CNEs are recognized, which greatly restricts the potential application of NLRs.

RESULTS

We propose a deep learning algorithm called ProNEP to identify NLR-effector pairs in a high-throughput manner. Specifically, we conceptualized the CNE prediction task as a protein-protein interaction (PPI) prediction task. Then, ProNEP predicts the interaction between NLRs and effectors by combining the transfer learning with a bilinear attention network. ProNEP achieves superior performance against state-of-the-art models designed for PPI predictions. Based on ProNEP, we conduct extensive identification of potential CNEs for 91 291 NLRs. With the rapid accumulation of genomic data, we expect that this tool will be widely used to predict CNEs in new species, advancing biology, immunology, and breeding.

AVAILABILITY AND IMPLEMENTATION

The ProNEP is available at http://nerrd.cn/#/prediction. The project code is available at https://github.com/QiaoYJYJ/ProNEP.

A quarantine paradox: understanding Gender-Based Violence (GBV) in post-COVID-19 era: insights from Golden Valley mining community, Zimbabwe

The COVID-19 pandemic has brought unprecedented challenges globally, with a notable surge in gender-based violence (GBV) incidents. This descriptive, exploratory study conducted in the Golden Valley mining community in Kadoma, Mashonaland West Province, Zimbabwe, delves into the challenges faced by GBV survivors during and after the pandemic, alongside community perceptions of GBV in the post-COVID-19 era. Guided by Bronfenbrenner's Social-Ecological Model which offers insights into the various levels of influence on GBV behavior and experiences,aiding in the development of contextually relevant prevention strategies. The study used qualitative methods such as interviews and focus group discussions, there were 24 study participant classified by random and convenient sampling techniques including traditional and community leaders, gbv surviours medical personnel and male and female, community members. The research followed the COREQ guidelines to transparently document the research process and findings. The research sheds light on the multifaceted nature of GBV exacerbated by societal norms and economic instability in a low-income, masculine-dominated work culture experience in the COVID-19 period. Findings underscore the urgent need for comprehensive prevention programs, effective legislation, and community engagement to address GBV in post-pandemic communities. The research provides valuable lessons for enhancing GBV prevention efforts globally, emphasizing the significance of survivor voices and addressing root causes of GBV.

The relational bottleneck as an inductive bias for efficient abstraction

A central challenge for cognitive science is to explain how abstract concepts are acquired from limited experience. This has often been framed in terms of a dichotomy between connectionist and symbolic cognitive models. Here, we highlight a recently emerging line of work that suggests a novel reconciliation of these approaches, by exploiting an inductive bias that we term the relational bottleneck. In that approach, neural networks are constrained via their architecture to focus on relations between perceptual inputs, rather than the attributes of individual inputs. We review a family of models that employ this approach to induce abstractions in a data-efficient manner, emphasizing their potential as candidate models for the acquisition of abstract concepts in the human mind and brain.

Digital Twin Studies for Reverse Engineering the Origins of Visual Intelligence

What are the core learning algorithms in brains? Nativists propose that intelligence emerges from innate domain-specific knowledge systems, whereas empiricists propose that intelligence emerges from domain-general systems that learn domain-specific knowledge from experience. We address this debate by reviewing digital twin studies designed to reverse engineer the learning algorithms in newborn brains. In digital twin studies, newborn animals and artificial agents are raised in the same environments and tested with the same tasks, permitting direct comparison of their learning abilities. Supporting empiricism, digital twin studies show that domain-general algorithms learn animal-like object perception when trained on the first-person visual experiences of newborn animals. Supporting nativism, digital twin studies show that domain-general algorithms produce innate domain-specific knowledge when trained on prenatal experiences (retinal waves). We argue that learning across humans, animals, and machines can be explained by a universal principle, which we call space-time fitting. Space-time fitting explains both empiricist and nativist phenomena, providing a unified framework for understanding the origins of intelligence.

Building Human-Like Artificial Agents: A General Cognitive Algorithm for Emulating Human Decision-Making in Dynamic Environments

One of the early goals of artificial intelligence (AI) was to create algorithms that exhibited behavior indistinguishable from human behavior (i.e., human-like behavior). Today, AI has diverged, often aiming to excel in tasks inspired by human capabilities and outperform humans, rather than replicating human cogntion and action. In this paper, I explore the overarching question of whether computational algorithms have achieved this initial goal of AI. I focus on dynamic decision-making, approaching the question from the perspective of computational cognitive science. I present a general cognitive algorithm that intends to emulate human decision-making in dynamic environments, as defined in instance-based learning theory (IBLT). I use the cognitive steps proposed in IBLT to organize and discuss current evidence that supports some of the human-likeness of the decision-making mechanisms. I also highlight the significant gaps in research that are required to improve current models and to create higher fidelity in computational algorithms to represent human decision processes. I conclude with concrete steps toward advancing the construction of algorithms that exhibit human-like behavior with the ultimate goal of supporting human dynamic decision-making.

Performance of GPT-4 with Vision on Text- and Image-based ACR Diagnostic Radiology In-Training Examination Questions

Background Recent advancements, including image processing capabilities, present new potential applications of large language models such as ChatGPT (OpenAI), a generative pretrained transformer, in radiology. However, baseline performance of ChatGPT in radiology-related tasks is understudied. Purpose To evaluate the performance of GPT-4 with vision (GPT-4V) on radiology in-training examination questions, including those with images, to gauge the model's baseline knowledge in radiology. Materials and Methods In this prospective study, conducted between September 2023 and March 2024, the September 2023 release of GPT-4V was assessed using 386 retired questions (189 image-based and 197 text-only questions) from the American College of Radiology Diagnostic Radiology In-Training Examinations. Nine question pairs were identified as duplicates; only the first instance of each duplicate was considered in ChatGPT's assessment. A subanalysis assessed the impact of different zero-shot prompts on performance. Statistical analysis included χ2 tests of independence to ascertain whether the performance of GPT-4V varied between question types or subspecialty. The McNemar test was used to evaluate performance differences between the prompts, with Benjamin-Hochberg adjustment of the P values conducted to control the false discovery rate (FDR). A P value threshold of less than.05 denoted statistical significance. Results GPT-4V correctly answered 246 (65.3%) of the 377 unique questions, with significantly higher accuracy on text-only questions (81.5%, 159 of 195) than on image-based questions (47.8%, 87 of 182) (χ2 test, P < .001). Subanalysis revealed differences between prompts on text-based questions, where chain-of-thought prompting outperformed long instruction by 6.1% (McNemar, P = .02; FDR = 0.063), basic prompting by 6.8% (P = .009, FDR = 0.044), and the original prompting style by 8.9% (P = .001, FDR = 0.014). No differences were observed between prompts on image-based questions with P values of .27 to >.99. Conclusion While GPT-4V demonstrated a level of competence in text-based questions, it showed deficits interpreting radiologic images. © RSNA, 2024 See also the editorial by Deng in this issue.

Psychological AI: Designing Algorithms Informed by Human Psychology

Psychological artificial intelligence (AI) applies insights from psychology to design computer algorithms. Its core domain is decision-making under uncertainty, that is, ill-defined situations that can change in unexpected ways rather than well-defined, stable problems, such as chess and Go. Psychological theories about heuristic processes under uncertainty can provide possible insights. I provide two illustrations. The first shows how recency-the human tendency to rely on the most recent information and ignore base rates-can be built into a simple algorithm that predicts the flu substantially better than did Google Flu Trends's big-data algorithms. The second uses a result from memory research-the paradoxical effect that making numbers less precise increases recall-in the design of algorithms that predict recidivism. These case studies provide an existence proof that psychological AI can help design efficient and transparent algorithms.

The Limitations of Large Language Models for Understanding Human Language and Cognition

Researchers have recently argued that the capabilities of Large Language Models (LLMs) can provide new insights into longstanding debates about the role of learning and/or innateness in the development and evolution of human language. Here, we argue on two grounds that LLMs alone tell us very little about human language and cognition in terms of acquisition and evolution. First, any similarities between human language and the output of LLMs are purely functional. Borrowing the "four questions" framework from ethology, we argue that what LLMs do is superficially similar, but how they do it is not. In contrast to the rich multimodal data humans leverage in interactive language learning, LLMs rely on immersive exposure to vastly greater quantities of unimodal text data, with recent multimodal efforts built upon mappings between images and text. Second, turning to functional similarities between human language and LLM output, we show that human linguistic behavior is much broader. LLMs were designed to imitate the very specific behavior of human writing; while they do this impressively, the underlying mechanisms of these models limit their capacities for meaning and naturalistic interaction, and their potential for dealing with the diversity in human language. We conclude by emphasising that LLMs are not theories of language, but tools that may be used to study language, and that can only be effectively applied with specific hypotheses to motivate research.

Strong and weak alignment of large language models with human values

Minimizing negative impacts of Artificial Intelligent (AI) systems on human societies without human supervision requires them to be able to align with human values. However, most current work only addresses this issue from a technical point of view, e.g., improving current methods relying on reinforcement learning from human feedback, neglecting what it means and is required for alignment to occur. Here, we propose to distinguish strong and weak value alignment. Strong alignment requires cognitive abilities (either human-like or different from humans) such as understanding and reasoning about agents' intentions and their ability to causally produce desired effects. We argue that this is required for AI systems like large language models (LLMs) to be able to recognize situations presenting a risk that human values may be flouted. To illustrate this distinction, we present a series of prompts showing ChatGPT's, Gemini's and Copilot's failures to recognize some of these situations. We moreover analyze word embeddings to show that the nearest neighbors of some human values in LLMs differ from humans' semantic representations. We then propose a new thought experiment that we call "the Chinese room with a word transition dictionary", in extension of John Searle's famous proposal. We finally mention current promising research directions towards a weak alignment, which could produce statistically satisfying answers in a number of common situations, however so far without ensuring any truth value.

Symbolic metaprogram search improves learning efficiency and explains rule learning in humans

Throughout their lives, humans seem to learn a variety of rules for things like applying category labels, following procedures, and explaining causal relationships. These rules are often algorithmically rich but are nonetheless acquired with minimal data and computation. Symbolic models based on program learning successfully explain rule-learning in many domains, but performance degrades quickly as program complexity increases. It remains unclear how to scale symbolic rule-learning methods to model human performance in challenging domains. Here we show that symbolic search over the space of metaprograms-programs that revise programs-dramatically improves learning efficiency. On a behavioral benchmark of 100 algorithmically rich rules, this approach fits human learning more accurately than alternative models while also using orders of magnitude less search. The computation required to match median human performance is consistent with conservative estimates of human thinking time. Our results suggest that metaprogram-like representations may help human learners to efficiently acquire rules.

A functional contextual, observer-centric, quantum mechanical, and neuro-symbolic approach to solving the alignment problem of artificial general intelligence: safe AI through intersecting computational psychological neuroscience and LLM architecture for emergent theory of mind

There have been impressive advancements in the field of natural language processing (NLP) in recent years, largely driven by innovations in the development of transformer-based large language models (LLM) that utilize "attention." This approach employs masked self-attention to establish (via similarly) different positions of tokens (words) within an inputted sequence of tokens to compute the most appropriate response based on its training corpus. However, there is speculation as to whether this approach alone can be scaled up to develop emergent artificial general intelligence (AGI), and whether it can address the alignment of AGI values with human values (called the alignment problem). Some researchers exploring the alignment problem highlight three aspects that AGI (or AI) requires to help resolve this problem: (1) an interpretable values specification; (2) a utility function; and (3) a dynamic contextual account of behavior. Here, a neurosymbolic model is proposed to help resolve these issues of human value alignment in AI, which expands on the transformer-based model for NLP to incorporate symbolic reasoning that may allow AGI to incorporate perspective-taking reasoning (i.e., resolving the need for a dynamic contextual account of behavior through deictics) as defined by a multilevel evolutionary and neurobiological framework into a functional contextual post-Skinnerian model of human language called "Neurobiological and Natural Selection Relational Frame Theory" (N-Frame). It is argued that this approach may also help establish a comprehensible value scheme, a utility function by expanding the expected utility equation of behavioral economics to consider functional contextualism, and even an observer (or witness) centric model for consciousness. Evolution theory, subjective quantum mechanics, and neuroscience are further aimed to help explain consciousness, and possible implementation within an LLM through correspondence to an interface as suggested by N-Frame. This argument is supported by the computational level of hypergraphs, relational density clusters, a conscious quantum level defined by QBism, and real-world applied level (human user feedback). It is argued that this approach could enable AI to achieve consciousness and develop deictic perspective-taking abilities, thereby attaining human-level self-awareness, empathy, and compassion toward others. Importantly, this consciousness hypothesis can be directly tested with a significance of approximately 5-sigma significance (with a 1 in 3.5 million probability that any identified AI-conscious observations in the form of a collapsed wave form are due to chance factors) through double-slit intent-type experimentation and visualization procedures for derived perspective-taking relational frames. Ultimately, this could provide a solution to the alignment problem and contribute to the emergence of a theory of mind (ToM) within AI.

Bridging flexible goal-directed cognition and consciousness: The Goal-Aligning Representation Internal Manipulation theory

Goal-directed manipulation of internal representations is a key element of human flexible behaviour, while consciousness is commonly associated with higher-order cognition and human flexibility. Current perspectives have only partially linked these processes, thus preventing a clear understanding of how they jointly generate flexible cognition and behaviour. Moreover, these limitations prevent an effective exploitation of this knowledge for technological scopes. We propose a new theoretical perspective that extends our 'three-component theory of flexible cognition' toward higher-order cognition and consciousness, based on the systematic integration of key concepts from Cognitive Neuroscience and AI/Robotics. The theory proposes that the function of conscious processes is to support the alignment of representations with multi-level goals. This higher alignment leads to more flexible and effective behaviours. We analyse here our previous model of goal-directed flexible cognition (validated with more than 20 human populations) as a starting GARIM-inspired model. By bridging the main theories of consciousness and goal-directed behaviour, the theory has relevant implications for scientific and technological fields. In particular, it contributes to developing new experimental tasks and interpreting clinical evidence. Finally, it indicates directions for improving machine learning and robotics systems and for informing real-world applications (e.g., in digital-twin healthcare and roboethics).

Abstract representations emerge in human hippocampal neurons during inference

Humans have the remarkable cognitive capacity to rapidly adapt to changing environments. Central to this capacity is the ability to form high-level, abstract representations that take advantage of regularities in the world to support generalization1. However, little is known about how these representations are encoded in populations of neurons, how they emerge through learning and how they relate to behaviour2,3. Here we characterized the representational geometry of populations of neurons (single units) recorded in the hippocampus, amygdala, medial frontal cortex and ventral temporal cortex of neurosurgical patients performing an inferential reasoning task. We found that only the neural representations formed in the hippocampus simultaneously encode several task variables in an abstract, or disentangled, format. This representational geometry is uniquely observed after patients learn to perform inference, and consists of disentangled directly observable and discovered latent task variables. Learning to perform inference by trial and error or through verbal instructions led to the formation of hippocampal representations with similar geometric properties. The observed relation between representational format and inference behaviour suggests that abstract and disentangled representational geometries are important for complex cognition.

Eight challenges in developing theory of intelligence

A good theory of mathematical beauty is more practical than any current observation, as new predictions about physical reality can be self-consistently verified. This belief applies to the current status of understanding deep neural networks including large language models and even the biological intelligence. Toy models provide a metaphor of physical reality, allowing mathematically formulating the reality (i.e., the so-called theory), which can be updated as more conjectures are justified or refuted. One does not need to present all details in a model, but rather, more abstract models are constructed, as complex systems such as the brains or deep networks have many sloppy dimensions but much less stiff dimensions that strongly impact macroscopic observables. This type of bottom-up mechanistic modeling is still promising in the modern era of understanding the natural or artificial intelligence. Here, we shed light on eight challenges in developing theory of intelligence following this theoretical paradigm. Theses challenges are representation learning, generalization, adversarial robustness, continual learning, causal learning, internal model of the brain, next-token prediction, and the mechanics of subjective experience.

Open-Vocabulary Predictive World Models from Sensor Observations

Cognitive scientists believe that adaptable intelligent agents like humans perform spatial reasoning tasks by learned causal mental simulation. The problem of learning these simulations is called predictive world modeling. We present the first framework for a learning open-vocabulary predictive world model (OV-PWM) from sensor observations. The model is implemented through a hierarchical variational autoencoder (HVAE) capable of predicting diverse and accurate fully observed environments from accumulated partial observations. We show that the OV-PWM can model high-dimensional embedding maps of latent compositional embeddings representing sets of overlapping semantics inferable by sufficient similarity inference. The OV-PWM simplifies the prior two-stage closed-set PWM approach to the single-stage end-to-end learning method. CARLA simulator experiments show that the OV-PWM can learn compact latent representations and generate diverse and accurate worlds with fine details like road markings, achieving 69 mIoU over six query semantics on an urban evaluation sequence. We propose the OV-PWM as a versatile continual learning paradigm for providing spatio-semantic memory and learned internal simulation capabilities to future general-purpose mobile robots.

Early-emerging combinatorial thought: Human infants flexibly combine kind and quantity concepts

Combinatorial thought, or the ability to combine a finite set of concepts into a myriad of complex ideas and knowledge structures, is the key to the productivity of the human mind and underlies communication, science, technology, and art. Despite the importance of combinatorial thought for human cognition and culture, its developmental origins remain unknown. To address this, we tested whether 12-mo-old infants (N = 60), who cannot yet speak and only understand a handful of words, can combine quantity and kind concepts activated by verbal input. We proceeded in two steps: first, we taught infants two novel labels denoting quantity (e.g., "mize" for 1 item; "padu" for 2 items, Experiment 1). Then, we assessed whether they could combine quantity and kind concepts upon hearing complex expressions comprising their labels (e.g., "padu duck", Experiments 2-3). At test, infants viewed four different sets of objects (e.g., 1 duck, 2 ducks, 1 ball, 2 balls) while being presented with the target phrase (e.g., "padu duck") naming one of them (e.g., 2 ducks). They successfully retrieved and combined on-line the labeled concepts, as evidenced by increased looking to the named sets but not to distractor sets. Our results suggest that combinatorial processes for building complex representations are available by the end of the first year of life. The infant mind seems geared to integrate concepts in novel productive ways. This ability may be a precondition for deciphering the ambient language(s) and building abstract models of experience that enable fast and flexible learning.

A mathematical theory of relational generalization in transitive inference

Humans and animals routinely infer relations between different items or events and generalize these relations to novel combinations of items. This allows them to respond appropriately to radically novel circumstances and is fundamental to advanced cognition. However, how learning systems (including the brain) can implement the necessary inductive biases has been unclear. We investigated transitive inference (TI), a classic relational task paradigm in which subjects must learn a relation ([Formula: see text] and [Formula: see text]) and generalize it to new combinations of items ([Formula: see text]). Through mathematical analysis, we found that a broad range of biologically relevant learning models (e.g. gradient flow or ridge regression) perform TI successfully and recapitulate signature behavioral patterns long observed in living subjects. First, we found that models with item-wise additive representations automatically encode transitive relations. Second, for more general representations, a single scalar "conjunctivity factor" determines model behavior on TI and, further, the principle of norm minimization (a standard statistical inductive bias) enables models with fixed, partly conjunctive representations to generalize transitively. Finally, neural networks in the "rich regime," which enables representation learning and improves generalization on many tasks, unexpectedly show poor generalization and anomalous behavior on TI. We find that such networks implement a form of norm minimization (over hidden weights) that yields a local encoding mechanism lacking transitivity. Our findings show how minimal statistical learning principles give rise to a classical relational inductive bias (transitivity), explain empirically observed behaviors, and establish a formal approach to understanding the neural basis of relational abstraction.

Joint modeling of choices and reaction times based on Bayesian contextual behavioral control

In cognitive neuroscience and psychology, reaction times are an important behavioral measure. However, in instrumental learning and goal-directed decision making experiments, findings often rely only on choice probabilities from a value-based model, instead of reaction times. Recent advancements have shown that it is possible to connect value-based decision models with reaction time models. However, typically these models do not provide an integrated account of both value-based choices and reaction times, but simply link two types of models. Here, we propose a novel integrative joint model of both choices and reaction times by combining a computational account of Bayesian sequential decision making with a sampling procedure. This allows us to describe how internal uncertainty in the planning process shapes reaction time distributions. Specifically, we use a recent context-specific Bayesian forward planning model which we extend by a Markov chain Monte Carlo (MCMC) sampler to obtain both choices and reaction times. As we will show this makes the sampler an integral part of the decision making process and enables us to reproduce, using simulations, well-known experimental findings in value based-decision making as well as classical inhibition and switching tasks. Specifically, we use the proposed model to explain both choice behavior and reaction times in instrumental learning and automatized behavior, in the Eriksen flanker task and in task switching. These findings show that the proposed joint behavioral model may describe common underlying processes in these different decision making paradigms.

Computational role of structure in neural activity and connectivity

One major challenge of neuroscience is identifying structure in seemingly disorganized neural activity. Different types of structure have different computational implications that can help neuroscientists understand the functional role of a particular brain area. Here, we outline a unified approach to characterize structure by inspecting the representational geometry and the modularity properties of the recorded activity and show that a similar approach can also reveal structure in connectivity. We start by setting up a general framework for determining geometry and modularity in activity and connectivity and relating these properties with computations performed by the network. We then use this framework to review the types of structure found in recent studies of model networks performing three classes of computations.

A sensory-motor theory of the neocortex

Recent neurophysiological and neuroanatomical studies suggest a close interaction between sensory and motor processes across the neocortex. Here, I propose that the neocortex implements active predictive coding (APC): each cortical area estimates both latent sensory states and actions (including potentially abstract actions internal to the cortex), and the cortex as a whole predicts the consequences of actions at multiple hierarchical levels. Feedback from higher areas modulates the dynamics of state and action networks in lower areas. I show how the same APC architecture can explain (1) how we recognize an object and its parts using eye movements, (2) why perception seems stable despite eye movements, (3) how we learn compositional representations, for example, part-whole hierarchies, (4) how complex actions can be planned using simpler actions, and (5) how we form episodic memories of sensory-motor experiences and learn abstract concepts such as a family tree. I postulate a mapping of the APC model to the laminar architecture of the cortex and suggest possible roles for cortico-cortical and cortico-subcortical pathways.

Flexible multitask computation in recurrent networks utilizes shared dynamical motifs

Flexible computation is a hallmark of intelligent behavior. However, little is known about how neural networks contextually reconfigure for different computations. In the present work, we identified an algorithmic neural substrate for modular computation through the study of multitasking artificial recurrent neural networks. Dynamical systems analyses revealed learned computational strategies mirroring the modular subtask structure of the training task set. Dynamical motifs, which are recurring patterns of neural activity that implement specific computations through dynamics, such as attractors, decision boundaries and rotations, were reused across tasks. For example, tasks requiring memory of a continuous circular variable repurposed the same ring attractor. We showed that dynamical motifs were implemented by clusters of units when the unit activation function was restricted to be positive. Cluster lesions caused modular performance deficits. Motifs were reconfigured for fast transfer learning after an initial phase of learning. This work establishes dynamical motifs as a fundamental unit of compositional computation, intermediate between neuron and network. As whole-brain studies simultaneously record activity from multiple specialized systems, the dynamical motif framework will guide questions about specialization and generalization.

A matched case-control analysis of autonomous vs human-driven vehicle accidents

Despite the recent advancements that Autonomous Vehicles have shown in their potential to improve safety and operation, considering differences between Autonomous Vehicles and Human-Driven Vehicles in accidents remain unidentified due to the scarcity of real-world Autonomous Vehicles accident data. We investigated the difference in accident occurrence between Autonomous Vehicles' levels and Human-Driven Vehicles by utilizing 2100 Advanced Driving Systems and Advanced Driver Assistance Systems and 35,113 Human-Driven Vehicles accident data. A matched case-control design was conducted to investigate the differential characteristics involving Autonomous' versus Human-Driven Vehicles' accidents. The analysis suggests that accidents of vehicles equipped with Advanced Driving Systems generally have a lower chance of occurring than Human-Driven Vehicles in most of the similar accident scenarios. However, accidents involving Advanced Driving Systems occur more frequently than Human-Driven Vehicle accidents under dawn/dusk or turning conditions, which is 5.25 and 1.98 times higher, respectively. Our research reveals the accident risk disparities between Autonomous Vehicles and Human-Driven Vehicles, informing future development in Autonomous technology and safety enhancements.

Reducing overconfident errors in molecular property classification using Posterior Network

Deep-learning-based classification models are increasingly used for predicting molecular properties in drug development. However, traditional classification models using the Softmax function often give overconfident mispredictions for out-of-distribution samples, highlighting a critical lack of accurate uncertainty estimation. Such limitations can result in substantial costs and should be avoided during drug development. Inspired by advances in evidential deep learning and Posterior Network, we replaced the Softmax function with a normalizing flow to enhance the uncertainty estimation ability of the model in molecular property classification. The proposed strategy was evaluated across diverse scenarios, including simulated experiments based on a synthetic dataset, ADMET predictions, and ligand-based virtual screening. The results demonstrate that compared with the vanilla model, the proposed strategy effectively alleviates the problem of giving overconfident but incorrect predictions. Our findings support the promising application of evidential deep learning in drug development and offer a valuable framework for further research.

A single fast Hebbian-like process enabling one-shot class addition in deep neural networks without backbone modification

One-shot learning, the ability to learn a new concept from a single instance, is a distinctive brain function that has garnered substantial interest in machine learning. While modeling physiological mechanisms poses challenges, advancements in artificial neural networks have led to performances in specific tasks that rival human capabilities. Proposing one-shot learning methods with these advancements, especially those involving simple mechanisms, not only enhance technological development but also contribute to neuroscience by proposing functionally valid hypotheses. Among the simplest methods for one-shot class addition with deep learning image classifiers is "weight imprinting," which uses neural activity from a new class image data as the corresponding new synaptic weights. Despite its simplicity, its relevance to neuroscience is ambiguous, and it often interferes with original image classification, which is a significant drawback in practical applications. This study introduces a novel interpretation where a part of the weight imprinting process aligns with the Hebbian rule. We show that a single Hebbian-like process enables pre-trained deep learning image classifiers to perform one-shot class addition without any modification to the original classifier's backbone. Using non-parametric normalization to mimic brain's fast Hebbian plasticity significantly reduces the interference observed in previous methods. Our method is one of the simplest and most practical for one-shot class addition tasks, and its reliance on a single fast Hebbian-like process contributes valuable insights to neuroscience hypotheses.

Socially adaptive cognitive architecture for human-robot collaboration in industrial settings

This paper introduces DAC-HRC, a novel cognitive architecture designed to optimize human-robot collaboration (HRC) in industrial settings, particularly within the context of Industry 4.0. The architecture is grounded in the Distributed Adaptive Control theory and the principles of joint intentionality and interdependence, which are key to effective HRC. Joint intentionality refers to the shared goals and mutual understanding between a human and a robot, while interdependence emphasizes the reliance on each other's capabilities to complete tasks. DAC-HRC is applied to a hybrid recycling plant for the disassembly and recycling of Waste Electrical and Electronic Equipment (WEEE) devices. The architecture incorporates several cognitive modules operating at different timescales and abstraction levels, fostering adaptive collaboration that is personalized to each human user. The effectiveness of DAC-HRC is demonstrated through several pilot studies, showcasing functionalities such as turn-taking interaction, personalized error-handling mechanisms, adaptive safety measures, and gesture-based communication. These features enhance human-robot collaboration in the recycling plant by promoting real-time robot adaptation to human needs and preferences. The DAC-HRC architecture aims to contribute to the development of a new HRC paradigm by paving the way for more seamless and efficient collaboration in Industry 4.0 by relying on socially adept cognitive architectures.

Compositionality in perception: A framework

Perception involves the processing of content or information about the world. In what form is this content represented? I argue that perception is widely compositional. The perceptual system represents many stimulus features (including shape, orientation, and motion) in terms of combinations of other features (such as shape parts, slant and tilt, common and residual motion vectors). But compositionality can take a variety of forms. The ways in which perceptual representations compose are markedly different from the ways in which sentences or thoughts are thought to be composed. I suggest that the thesis that perception is compositional is not itself a concrete hypothesis with specific predictions; rather it affords a productive framework for developing and evaluating specific empirical hypotheses about the form and content of perceptual representations. The question is not just whether perception is compositional, but how. Answering this latter question can provide fundamental insights into perception. This article is categorized under: Philosophy > Representation Philosophy > Foundations of Cognitive Science Psychology > Perception and Psychophysics.

Invariant representations in abstract concept grounding - the physical world in grounded cognition

Grounded cognition states that mental representations of concepts consist of experiential aspects. For example, the concept "cup" consists of the sensorimotor experiences from interactions with cups. Typical modalities in which concepts are grounded are: The sensorimotor system (including interoception), emotion, action, language, and social aspects. Here, we argue that this list should be expanded to include physical invariants (unchanging features of physical motion; e.g., gravity, momentum, friction). Research on physical reasoning consistently demonstrates that physical invariants are represented as fundamentally as other grounding substrates, and therefore should qualify. We assess several theories of concept representation (simulation, conceptual metaphor, conceptual spaces, predictive processing) and their positions on physical invariants. We find that the classic grounded cognition theories, simulation and conceptual metaphor theory, have not considered physical invariants, while conceptual spaces and predictive processing have. We conclude that physical invariants should be included into grounded cognition theories, and that the core mechanisms of simulation and conceptual metaphor theory are well suited to do this. Furthermore, conceptual spaces and predictive processing are very promising and should also be integrated with grounded cognition in the future.

Linear Codes for Hyperdimensional Computing

Hyperdimensional computing (HDC) is an emerging computational paradigm for representing compositional information as high-dimensional vectors and has a promising potential in applications ranging from machine learning to neuromorphic computing. One of the long-standing challenges in HDC is factoring a compositional representation to its constituent factors, also known as the recovery problem. In this article, we take a novel approach to solve the recovery problem and propose the use of random linear codes. These codes are subspaces over the Boolean field and are a well-studied topic in information theory with various applications in digital communication. We begin by showing that hyperdimensional encoding using random linear codes retains favorable properties of the prevalent (ordinary) random codes; hence, HD representations using the two methods have comparable information storage capabilities. We proceed to show that random linear codes offer a rich subcode structure that can be used to form key-value stores, which encapsulate the most used cases of HDC. Most important, we show that under the framework we develop, random linear codes admit simple recovery algorithms to factor (either bundled or bound) compositional representations. The former relies on constructing certain linear equation systems over the Boolean field, the solution to which reduces the search space dramatically and strictly outperforms exhaustive search in many cases. The latter employs the subspace structure of these codes to achieve provably correct factorization. Both methods are strictly faster than the state-of-the-art resonator networks, often by an order of magnitude. We implemented our techniques in Python using a benchmark software library and demonstrated promising experimental results.

Toward the Emergence of Intelligent Control: Episodic Generalization and Optimization

Human cognition is unique in its ability to perform a wide range of tasks and to learn new tasks quickly. Both abilities have long been associated with the acquisition of knowledge that can generalize across tasks and the flexible use of that knowledge to execute goal-directed behavior. We investigate how this emerges in a neural network by describing and testing the Episodic Generalization and Optimization (EGO) framework. The framework consists of an episodic memory module, which rapidly learns relationships between stimuli; a semantic pathway, which more slowly learns how stimuli map to responses; and a recurrent context module, which maintains a representation of task-relevant context information, integrates this over time, and uses it both to recall context-relevant memories (in episodic memory) and to bias processing in favor of context-relevant features and responses (in the semantic pathway). We use the framework to address empirical phenomena across reinforcement learning, event segmentation, and category learning, showing in simulations that the same set of underlying mechanisms accounts for human performance in all three domains. The results demonstrate how the components of the EGO framework can efficiently learn knowledge that can be flexibly generalized across tasks, furthering our understanding of how humans can quickly learn how to perform a wide range of tasks-a capability that is fundamental to human intelligence.

A stochastic world model on gravity for stability inference

The fact that objects without proper support will fall to the ground is not only a natural phenomenon, but also common sense in mind. Previous studies suggest that humans may infer objects' stability through a world model that performs mental simulations with a priori knowledge of gravity acting upon the objects. Here we measured participants' sensitivity to gravity to investigate how the world model works. We found that the world model on gravity was not a faithful replica of the physical laws, but instead encoded gravity's vertical direction as a Gaussian distribution. The world model with this stochastic feature fit nicely with participants' subjective sense of objects' stability and explained the illusion that taller objects are perceived as more likely to fall. Furthermore, a computational model with reinforcement learning revealed that the stochastic characteristic likely originated from experience-dependent comparisons between predictions formed by internal simulations and the realities observed in the external world, which illustrated the ecological advantage of stochastic representation in balancing accuracy and speed for efficient stability inference. The stochastic world model on gravity provides an example of how a priori knowledge of the physical world is implemented in mind that helps humans operate flexibly in open-ended environments.

From task structures to world models: what do LLMs know?

In what sense does a large language model (LLM) have knowledge? We answer by granting LLMs 'instrumental knowledge': knowledge gained by using next-word generation as an instrument. We then ask how instrumental knowledge is related to the ordinary, 'worldly knowledge' exhibited by humans, and explore this question in terms of the degree to which instrumental knowledge can be said to incorporate the structured world models of cognitive science. We discuss ways LLMs could recover degrees of worldly knowledge and suggest that such recovery will be governed by an implicit, resource-rational tradeoff between world models and tasks. Our answer to this question extends beyond the capabilities of a particular AI system and challenges assumptions about the nature of knowledge and intelligence.

Unsupervised Object-Centric Learning From Multiple Unspecified Viewpoints

Visual scenes are extremely diverse, not only because there are infinite possible combinations of objects and backgrounds but also because the observations of the same scene may vary greatly with the change of viewpoints. When observing a multi-object visual scene from multiple viewpoints, humans can perceive the scene compositionally from each viewpoint while achieving the so-called "object constancy" across different viewpoints, even though the exact viewpoints are untold. This ability is essential for humans to identify the same object while moving and to learn from vision efficiently. It is intriguing to design models that have a similar ability. In this article, we consider a novel problem of learning compositional scene representations from multiple unspecified (i.e., unknown and unrelated) viewpoints without using any supervision and propose a deep generative model which separates latent representations into a viewpoint-independent part and a viewpoint-dependent part to solve this problem. During the inference, latent representations are randomly initialized and iteratively updated by integrating the information in different viewpoints with neural networks. Experiments on several specifically designed synthetic datasets have shown that the proposed method can effectively learn from multiple unspecified viewpoints.

Toward Improving the Generation Quality of Autoregressive Slot VAEs

Unconditional scene inference and generation are challenging to learn jointly with a single compositional model. Despite encouraging progress on models that extract object-centric representations ("slots") from images, unconditional generation of scenes from slots has received less attention. This is primarily because learning the multiobject relations necessary to imagine coherent scenes is difficult. We hypothesize that most existing slot-based models have a limited ability to learn object correlations. We propose two improvements that strengthen object correlation learning. The first is to condition the slots on a global, scene-level variable that captures higher-order correlations between slots. Second, we address the fundamental lack of a canonical order for objects in images by proposing to learn a consistent order to use for the autoregressive generation of scene objects. Specifically, we train an autoregressive slot prior to sequentially generate scene objects following a learned order. Ordered slot inference entails first estimating a randomly ordered set of slots using existing approaches for extracting slots from images, then aligning those slots to ordered slots generated autoregressively with the slot prior. Our experiments across three multiobject environments demonstrate clear gains in unconditional scene generation quality. Detailed ablation studies are also provided that validate the two proposed improvements.

Human-level few-shot concept induction through minimax entropy learning

Humans learn concepts both from labeled supervision and by unsupervised observation of patterns, a process machines are being taught to mimic by training on large annotated datasets-a method quite different from the human pathway, wherein few examples with no supervision suffice to induce an unfamiliar relational concept. We introduce a computational model designed to emulate human inductive reasoning on abstract reasoning tasks, such as those in IQ tests, using a minimax entropy approach. This method combines identifying the most effective constraints on data via minimum entropy with determining the best combination of them via maximum entropy. Our model, which applies this unsupervised technique, induces concepts from just one instance, reaching human-level performance on tasks of Raven's Progressive Matrices (RPM), Machine Number Sense (MNS), and Odd-One-Out (O3). These results demonstrate the potential of minimax entropy learning for enabling machines to learn relational concepts efficiently with minimal input.

Bayesian Reinforcement Learning With Limited Cognitive Load

All biological and artificial agents must act given limits on their ability to acquire and process information. As such, a general theory of adaptive behavior should be able to account for the complex interactions between an agent's learning history, decisions, and capacity constraints. Recent work in computer science has begun to clarify the principles that shape these dynamics by bridging ideas from reinforcement learning, Bayesian decision-making, and rate-distortion theory. This body of work provides an account of capacity-limited Bayesian reinforcement learning, a unifying normative framework for modeling the effect of processing constraints on learning and action selection. Here, we provide an accessible review of recent algorithms and theoretical results in this setting, paying special attention to how these ideas can be applied to studying questions in the cognitive and behavioral sciences.

Neural dynamic foundations of a theory of higher cognition: the case of grounding nested phrases

Because cognitive competences emerge in evolution and development from the sensory-motor domain, we seek a neural process account for higher cognition in which all representations are necessarily grounded in perception and action. The challenge is to understand how hallmarks of higher cognition, productivity, systematicity, and compositionality, may emerge from such a bottom-up approach. To address this challenge, we present key ideas from Dynamic Field Theory which postulates that neural populations are organized by recurrent connectivity to create stable localist representations. Dynamic instabilities enable the autonomous generation of sequences of mental states. The capacity to apply neural circuitry across broad sets of inputs that emulates the function call postulated in symbolic computation emerges through coordinate transforms implemented in neural gain fields. We show how binding localist neural representations through a shared index dimension enables conceptual structure, in which the interdependence among components of a representation is flexibly expressed. We demonstrate these principles in a neural dynamic architecture that represents and perceptually grounds nested relational and action phrases. Sequences of neural processing steps are generated autonomously to attentionally select the referenced objects and events in a manner that is sensitive to their interdependencies. This solves the problem of 2 and the massive binding problem in expressions such as "the small tree that is to the left of the lake which is to the left of the large tree". We extend earlier work by incorporating new types of grammatical constructions and a larger vocabulary. We discuss the DFT framework relative to other neural process accounts of higher cognition and assess the scope and challenges of such neural theories.

Stochastic neuro-fuzzy system implemented in memristor crossbar arrays

Neuro-symbolic artificial intelligence has garnered considerable attention amid increasing industry demands for high-performance neural networks that are interpretable and adaptable to previously unknown problem domains with minimal reconfiguration. However, implementing neuro-symbolic hardware is challenging due to the complexity in symbolic knowledge representation and calculation. We experimentally demonstrated a memristor-based neuro-fuzzy hardware based on TiN/TaOx/HfOx/TiN chips that is superior to its silicon-based counterpart in terms of throughput and energy efficiency by using array topological structure for knowledge representation and physical laws for computing. Intrinsic memristor variability is fully exploited to increase robustness in knowledge representation. A hybrid in situ training strategy is proposed for error minimizing in training. The hardware adapts easier to a previously unknown environment, achieving ~6.6 times faster convergence and ~6 times lower error than deep learning. The hardware energy efficiency is over two orders of magnitude greater than field-programmable gate arrays. This research greatly extends the capability of memristor-based neuromorphic computing systems in artificial intelligence.

Few-Shot Learning With a Strong Teacher

Few-shot learning (FSL) aims to generate a classifier using limited labeled examples. Many existing works take the meta-learning approach, constructing a few-shot learner (a meta-model) that can learn from few-shot examples to generate a classifier. Typically, the few-shot learner is constructed or meta-trained by sampling multiple few-shot tasks in turn and optimizing the few-shot learner's performance in generating classifiers for those tasks. The performance is measured by how well the resulting classifiers classify the test (i.e., query) examples of those tasks. In this paper, we point out two potential weaknesses of this approach. First, the sampled query examples may not provide sufficient supervision for meta-training the few-shot learner. Second, the effectiveness of meta-learning diminishes sharply with the increasing number of shots (i.e., the number of training examples per class). To resolve these issues, we propose a novel meta-training objective for the few-shot learner, which is to encourage the few-shot learner to generate classifiers that perform like strong classifiers. Concretely, we associate each sampled few-shot task with a strong classifier, which is trained with ample labeled examples. The strong classifiers can be seen as the target classifiers that we hope the few-shot learner to generate given few-shot examples, and we use the strong classifiers to supervise the few-shot learner. We present an efficient way to construct the strong classifier, making our proposed objective an easily plug-and-play term to existing meta-learning based FSL methods. We validate our approach, (Learning with A Strong Teacher for few-SHOT learning), in combinations with many representative meta-learning methods. On several benchmark datasets including miniImageNet and tieredImageNet, our approach leads to a notable improvement across a variety of tasks. More importantly, with our approach, meta-learning based FSL methods can consistently outperform non-meta-learning based methods at different numbers of shots, even in many-shot settings, greatly strengthening their applicability.

The tri-flow adaptiveness of codes in major evolutionary transitions

Life codes increase in both number and variety with biological complexity. Although our knowledge of codes is constantly expanding, the evolutionary progression of organic, neural, and cultural codes in response to selection pressure remains poorly understood. Greater clarification of the selective mechanisms is achieved by investigating how major evolutionary transitions reduce spatiotemporal and energetic constraints on transmitting heritable code to offspring. Evolution toward less constrained flows is integral to enduring flow architecture everywhere, in both engineered and natural flow systems. Beginning approximately 4 billion years ago, the most basic level for transmitting genetic material to offspring was initiated by protocell division. Evidence from ribosomes suggests that protocells transmitted comma-free or circular codes, preceding the evolution of standard genetic code. This rudimentary information flow within protocells is likely to have first emerged within the geo-energetic and geospatial constraints of hydrothermal vents. A broad-gauged hypothesis is that major evolutionary transitions overcame such constraints with tri-flow adaptations. The interconnected triple flows incorporated energy-converting, spatiotemporal, and code-based informational dynamics. Such tri-flow adaptations stacked sequence splicing code on top of protein-DNA recognition code in eukaryotes, prefiguring the transition to sexual reproduction. Sex overcame the spatiotemporal-energetic constraints of binary fission with further code stacking. Examples are tubulin code and transcription initiation code in vertebrates. In a later evolutionary transition, language reduced metabolic-spatiotemporal constraints on inheritance by stacking phonetic, phonological, and orthographic codes. In organisms that reproduce sexually, each major evolutionary transition is shown to be a tri-flow adaptation that adds new levels of code-based informational exchange. Evolving biological complexity is also shown to increase the nongenetic transmissibility of code.

Top-down generation of low-resolution representations improves visual perception and imagination

Perception or imagination requires top-down signals from high-level cortex to primary visual cortex (V1) to reconstruct or simulate the representations bottom-up stimulated by the seen images. Interestingly, top-down signals in V1 have lower spatial resolution than bottom-up representations. It is unclear why the brain uses low-resolution signals to reconstruct or simulate high-resolution representations. By modeling the top-down pathway of the visual system using the decoder of a variational auto-encoder (VAE), we reveal that low-resolution top-down signals can better reconstruct or simulate the information contained in the sparse activities of V1 simple cells, which facilitates perception and imagination. This advantage of low-resolution generation is related to facilitating high-level cortex to form geometry-respecting representations observed in experiments. Furthermore, we present two findings regarding this phenomenon in the context of AI-generated sketches, a style of drawings made of lines. First, we found that the quality of the generated sketches critically depends on the thickness of the lines in the sketches: thin-line sketches are harder to generate than thick-line sketches. Second, we propose a technique to generate high-quality thin-line sketches: instead of directly using original thin-line sketches, we use blurred sketches to train VAE or GAN (generative adversarial network), and then infer the thin-line sketches from the VAE- or GAN-generated blurred sketches. Collectively, our work suggests that low-resolution top-down generation is a strategy the brain uses to improve visual perception and imagination, which inspires new sketch-generation AI techniques.

Minimap: An interactive dynamic decision making game for search and rescue missions

Many aspects of humans' dynamic decision-making (DDM) behaviors have been studied with computer-simulated games called microworlds. However, most microworlds only emphasize specific elements of DDM and are inflexible in generating a variety of environments and experimental designs. Moreover, despite the ubiquity of gridworld games for Artificial Intelligence (AI) research, only some tools exist to aid in the development of browser-based gridworld environments for studying the dynamics of human decision-making behavior. To address these issues, we introduce Minimap, a dynamic interactive game to examine DDM in search and rescue missions, which incorporates all the essential characteristics of DDM and offers a wide range of flexibility regarding experimental setups and the creation of experimental scenarios. Minimap specifically allows customization of dynamics, complexity, opaqueness, and dynamic complexity when designing a DDM task. Minimap also enables researchers to visualize and replay recorded human trajectories for the analysis of human behavior. To demonstrate the utility of Minimap, we present a behavioral experiment that examines the impact of different degrees of structural complexity coupled with the opaqueness of the environment on human decision-making performance under time constraints. We discuss the potential applications of Minimap in improving productivity and transparent replications of human behavior and human-AI teaming research. We made Minimap an open-source tool, freely available at  https://github.com/DDM-Lab/MinimapInteractiveDDMGame .

AI in medical diagnosis: AI prediction & human judgment

AI has long been regarded as a panacea for decision-making and many other aspects of knowledge work; as something that will help humans get rid of their shortcomings. We believe that AI can be a useful asset to support decision-makers, but not that it should replace decision-makers. Decision-making uses algorithmic analysis, but it is not solely algorithmic analysis; it also involves other factors, many of which are very human, such as creativity, intuition, emotions, feelings, and value judgments. We have conducted semi-structured open-ended research interviews with 17 dermatologists to understand what they expect from an AI application to deliver to medical diagnosis. We have found four aggregate dimensions along which the thinking of dermatologists can be described: the ways in which our participants chose to interact with AI, responsibility, 'explainability', and the new way of thinking (mindset) needed for working with AI. We believe that our findings will help physicians who might consider using AI in their diagnosis to understand how to use AI beneficially. It will also be useful for AI vendors in improving their understanding of how medics want to use AI in diagnosis. Further research will be needed to examine if our findings have relevance in the wider medical field and beyond.

Adopting artificial intelligence in cardiovascular medicine: a scoping review

Recent years have witnessed significant transformations in cardiovascular medicine, driven by the rapid evolution of artificial intelligence (AI). This scoping review was conducted to capture the breadth of AI applications within cardiovascular science. Employing a structured approach, we sourced relevant articles from PubMed, with an emphasis on journals encompassing general cardiology and digital medicine. We applied filters to highlight cardiovascular articles published in journals focusing on general internal medicine, cardiology and digital medicine, thereby identifying the prevailing trends in the field. Following a comprehensive full-text screening, a total of 140 studies were identified. Over the preceding 5 years, cardiovascular medicine's interplay with AI has seen an over tenfold augmentation. This expansive growth encompasses multiple cardiovascular subspecialties, including but not limited to, general cardiology, ischemic heart disease, heart failure, and arrhythmia. Deep learning emerged as the predominant methodology. The majority of AI endeavors in this domain have been channeled toward enhancing diagnostic and prognostic capabilities, utilizing resources such as hospital datasets, electrocardiograms, and echocardiography. A significant uptrend was observed in AI's application for omics data analysis. However, a clear gap persists in AI's full-scale integration into the clinical decision-making framework. AI, particularly deep learning, has demonstrated robust applications across cardiovascular subspecialties, indicating its transformative potential in this field. As we continue on this trajectory, ensuring the alignment of technological progress with medical ethics becomes crucial. The abundant digital health data today further accentuates the need for meticulous systematic reviews, tailoring them to each cardiovascular subspecialty.

Systematic Human Learning and Generalization From a Brief Tutorial With Explanatory Feedback

We investigate human adults' ability to learn an abstract reasoning task quickly and to generalize outside of the range of training examples. Using a task based on a solution strategy in Sudoku, we provide Sudoku-naive participants with a brief instructional tutorial with explanatory feedback using a narrow range of training examples. We find that most participants who master the task do so within 10 practice trials and generalize well to puzzles outside of the training range. We also find that most of those who master the task can describe a valid solution strategy, and such participants perform better on transfer puzzles than those whose strategy descriptions are vague or incomplete. Interestingly, fewer than half of our human participants were successful in acquiring a valid solution strategy, and this ability was associated with completion of high school algebra and geometry. We consider the implications of these findings for understanding human systematic reasoning, as well as the challenges these findings pose for building computational models that capture all aspects of our findings, and we point toward a role for learning from instructions and explanations to support rapid learning and generalization.

Compositional diversity in visual concept learning

Humans leverage compositionality to efficiently learn new concepts, understanding how familiar parts can combine together to form novel objects. In contrast, popular computer vision models struggle to make the same types of inferences, requiring more data and generalizing less flexibly than people do. Here, we study these distinctively human abilities across a range of different types of visual composition, examining how people classify and generate "alien figures" with rich relational structure. We also develop a Bayesian program induction model which searches for the best programs for generating the candidate visual figures, utilizing a large program space containing different compositional mechanisms and abstractions. In few shot classification tasks, we find that people and the program induction model can make a range of meaningful compositional generalizations, with the model providing a strong account of the experimental data as well as interpretable parameters that reveal human assumptions about the factors invariant to category membership (here, to rotation and changing part attachment). In few shot generation tasks, both people and the models are able to construct compelling novel examples, with people behaving in additional structured ways beyond the model capabilities, e.g. making choices that complete a set or reconfigure existing parts in new ways. To capture these additional behavioral patterns, we develop an alternative model based on neuro-symbolic program induction: this model also composes new concepts from existing parts yet, distinctively, it utilizes neural network modules to capture residual statistical structure. Together, our behavioral and computational findings show how people and models can produce a variety of compositional behavior when classifying and generating visual objects.

Commonsense knowledge in cognitive robotics: a systematic literature review

One of the big challenges in robotics is the generalization necessary for performing unknown tasks in unknown environments on unknown objects. For us humans, this challenge is simplified by the commonsense knowledge we can access. For cognitive robotics, representing and acquiring commonsense knowledge is a relevant problem, so we perform a systematic literature review to investigate the current state of commonsense knowledge exploitation in cognitive robotics. For this review, we combine a keyword search on six search engines with a snowballing search on six related reviews, resulting in 2,048 distinct publications. After applying pre-defined inclusion and exclusion criteria, we analyse the remaining 52 publications. Our focus lies on the use cases and domains for which commonsense knowledge is employed, the commonsense aspects that are considered, the datasets/resources used as sources for commonsense knowledge and the methods for evaluating these approaches. Additionally, we discovered a divide in terminology between research from the knowledge representation and reasoning and the cognitive robotics community. This divide is investigated by looking at the extensive review performed by Zech et al. (The International Journal of Robotics Research, 2019, 38, 518-562), with whom we have no overlapping publications despite the similar goals.