Computational Foundations of Natural Intelligence

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.

Novel Domain Knowledge-Encoding Algorithm Enables Label-Efficient Deep Learning for Cardiac CT Segmentation to Guide Atrial Fibrillation Treatment in a Pilot Dataset

Background: Segmenting computed tomography (CT) is crucial in various clinical applications, such as tailoring personalized cardiac ablation for managing cardiac arrhythmias. Automating segmentation through machine learning (ML) is hindered by the necessity for large, labeled training data, which can be challenging to obtain. This article proposes a novel approach for automated, robust labeling using domain knowledge to achieve high-performance segmentation by ML from a small training set. The approach, the domain knowledge-encoding (DOKEN) algorithm, reduces the reliance on large training datasets by encoding cardiac geometry while automatically labeling the training set. The method was validated in a hold-out dataset of CT results from an atrial fibrillation (AF) ablation study. Methods: The DOKEN algorithm parses left atrial (LA) structures, extracts "anatomical knowledge" by leveraging digital LA models (available publicly), and then applies this knowledge to achieve high ML segmentation performance with a small number of training samples. The DOKEN-labeled training set was used to train a nnU-Net deep neural network (DNN) model for segmenting cardiac CT in N = 20 patients. Subsequently, the method was tested in a hold-out set with N = 100 patients (five times larger than training set) who underwent AF ablation. Results: The DOKEN algorithm integrated with the nn-Unet model achieved high segmentation performance with few training samples, with a training to test ratio of 1:5. The Dice score of the DOKEN-enhanced model was 96.7% (IQR: 95.3% to 97.7%), with a median error in surface distance of boundaries of 1.51 mm (IQR: 0.72 to 3.12) and a mean centroid-boundary distance of 1.16 mm (95% CI: -4.57 to 6.89), similar to expert results (r = 0.99; p < 0.001). In digital hearts, the novel DOKEN approach segmented the LA structures with a mean difference for the centroid-boundary distances of -0.27 mm (95% CI: -3.87 to 3.33; r = 0.99; p < 0.0001). Conclusions: The proposed novel domain knowledge-encoding algorithm was able to perform the segmentation of six substructures of the LA, reducing the need for large training data sets. The combination of domain knowledge encoding and a machine learning approach could reduce the dependence of ML on large training datasets and could potentially be applied to AF ablation procedures and extended in the future to other imaging, 3D printing, and data science applications.

Predictive coding and stochastic resonance as fundamental principles of auditory phantom perception

Mechanistic insight is achieved only when experiments are employed to test formal or computational models. Furthermore, in analogy to lesion studies, phantom perception may serve as a vehicle to understand the fundamental processing principles underlying healthy auditory perception. With a special focus on tinnitus-as the prime example of auditory phantom perception-we review recent work at the intersection of artificial intelligence, psychology and neuroscience. In particular, we discuss why everyone with tinnitus suffers from (at least hidden) hearing loss, but not everyone with hearing loss suffers from tinnitus. We argue that intrinsic neural noise is generated and amplified along the auditory pathway as a compensatory mechanism to restore normal hearing based on adaptive stochastic resonance. The neural noise increase can then be misinterpreted as auditory input and perceived as tinnitus. This mechanism can be formalized in the Bayesian brain framework, where the percept (posterior) assimilates a prior prediction (brain's expectations) and likelihood (bottom-up neural signal). A higher mean and lower variance (i.e. enhanced precision) of the likelihood shifts the posterior, evincing a misinterpretation of sensory evidence, which may be further confounded by plastic changes in the brain that underwrite prior predictions. Hence, two fundamental processing principles provide the most explanatory power for the emergence of auditory phantom perceptions: predictive coding as a top-down and adaptive stochastic resonance as a complementary bottom-up mechanism. We conclude that both principles also play a crucial role in healthy auditory perception. Finally, in the context of neuroscience-inspired artificial intelligence, both processing principles may serve to improve contemporary machine learning techniques.

Exploring weight initialization, diversity of solutions, and degradation in recurrent neural networks trained for temporal and decision-making tasks

Recurrent Neural Networks (RNNs) are frequently used to model aspects of brain function and structure. In this work, we trained small fully-connected RNNs to perform temporal and flow control tasks with time-varying stimuli. Our results show that different RNNs can solve the same task by converging to different underlying dynamics and also how the performance gracefully degrades as either network size is decreased, interval duration is increased, or connectivity damage is induced. For the considered tasks, we explored how robust the network obtained after training can be according to task parameterization. In the process, we developed a framework that can be useful to parameterize other tasks of interest in computational neuroscience. Our results are useful to quantify different aspects of the models, which are normally used as black boxes and need to be understood in order to model the biological response of cerebral cortex areas.

Artificial Collective Intelligence Engineering: A Survey of Concepts and Perspectives

Collectiveness is an important property of many systems-both natural and artificial. By exploiting a large number of individuals, it is often possible to produce effects that go far beyond the capabilities of the smartest individuals or even to produce intelligent collective behavior out of not-so-intelligent individuals. Indeed, collective intelligence, namely, the capability of a group to act collectively in a seemingly intelligent way, is increasingly often a design goal of engineered computational systems-motivated by recent technoscientific trends like the Internet of Things, swarm robotics, and crowd computing, to name only a few. For several years, the collective intelligence observed in natural and artificial systems has served as a source of inspiration for engineering ideas, models, and mechanisms. Today, artificial and computational collective intelligence are recognized research topics, spanning various techniques, kinds of target systems, and application domains. However, there is still a lot of fragmentation in the research panorama of the topic within computer science, and the verticality of most communities and contributions makes it difficult to extract the core underlying ideas and frames of reference. The challenge is to identify, place in a common structure, and ultimately connect the different areas and methods addressing intelligent collectives. To address this gap, this article considers a set of broad scoping questions providing a map of collective intelligence research, mostly by the point of view of computer scientists and engineers. Accordingly, it covers preliminary notions, fundamental concepts, and the main research perspectives, identifying opportunities and challenges for researchers on artificial and computational collective intelligence engineering.

Segmenting computed tomograms for cardiac ablation using machine learning leveraged by domain knowledge encoding

Background

Segmentation of computed tomography (CT) is important for many clinical procedures including personalized cardiac ablation for the management of cardiac arrhythmias. While segmentation can be automated by machine learning (ML), it is limited by the need for large, labeled training data that may be difficult to obtain. We set out to combine ML of cardiac CT with domain knowledge, which reduces the need for large training datasets by encoding cardiac geometry, which we then tested in independent datasets and in a prospective study of atrial fibrillation (AF) ablation.

Methods

We mathematically represented atrial anatomy with simple geometric shapes and derived a model to parse cardiac structures in a small set of N = 6 digital hearts. The model, termed "virtual dissection," was used to train ML to segment cardiac CT in N = 20 patients, then tested in independent datasets and in a prospective study.

Results

In independent test cohorts (N = 160) from 2 Institutions with different CT scanners, atrial structures were accurately segmented with Dice scores of 96.7% in internal (IQR: 95.3%-97.7%) and 93.5% in external (IQR: 91.9%-94.7%) test data, with good agreement with experts (r = 0.99; p < 0.0001). In a prospective study of 42 patients at ablation, this approach reduced segmentation time by 85% (2.3 ± 0.8 vs. 15.0 ± 6.9 min, p < 0.0001), yet provided similar Dice scores to experts (93.9% (IQR: 93.0%-94.6%) vs. 94.4% (IQR: 92.8%-95.7%), p = NS).

Conclusions

Encoding cardiac geometry using mathematical models greatly accelerated training of ML to segment CT, reducing the need for large training sets while retaining accuracy in independent test data. Combining ML with domain knowledge may have broad applications.

The neuroconnectionist research programme

Artificial neural networks (ANNs) inspired by biology are beginning to be widely used to model behavioural and neural data, an approach we call 'neuroconnectionism'. ANNs have been not only lauded as the current best models of information processing in the brain but also criticized for failing to account for basic cognitive functions. In this Perspective article, we propose that arguing about the successes and failures of a restricted set of current ANNs is the wrong approach to assess the promise of neuroconnectionism for brain science. Instead, we take inspiration from the philosophy of science, and in particular from Lakatos, who showed that the core of a scientific research programme is often not directly falsifiable but should be assessed by its capacity to generate novel insights. Following this view, we present neuroconnectionism as a general research programme centred around ANNs as a computational language for expressing falsifiable theories about brain computation. We describe the core of the programme, the underlying computational framework and its tools for testing specific neuroscientific hypotheses and deriving novel understanding. Taking a longitudinal view, we review past and present neuroconnectionist projects and their responses to challenges and argue that the research programme is highly progressive, generating new and otherwise unreachable insights into the workings of the brain.

Artificial Intelligence for Cancer Detection-A Bibliometric Analysis and Avenues for Future Research

After cardiovascular diseases, cancer is responsible for the most deaths worldwide. Detecting a cancer disease early improves the chances for healing significantly. One group of technologies that is increasingly applied for detecting cancer is artificial intelligence. Artificial intelligence has great potential to support clinicians and medical practitioners as it allows for the early detection of carcinomas. During recent years, research on artificial intelligence for cancer detection grew a lot. Within this article, we conducted a bibliometric study of the existing research dealing with the application of artificial intelligence in cancer detection. We analyzed 6450 articles on that topic that were published between 1986 and 2022. By doing so, we were able to give an overview of this research field, including its key topics, relevant outlets, institutions, and articles. Based on our findings, we developed a future research agenda that can help to advance research on artificial intelligence for cancer detection. In summary, our study is intended to serve as a platform and foundation for researchers that are interested in the potential of artificial intelligence for detecting cancer.

Using Artificial Intelligence for Drug Discovery: A Bibliometric Study and Future Research Agenda

Drug discovery is usually a rule-based process that is carefully carried out by pharmacists. However, a new trend is emerging in research and practice where artificial intelligence is being used for drug discovery to increase efficiency or to develop new drugs for previously untreatable diseases. Nevertheless, so far, no study takes a holistic view of AI-based drug discovery research. Given the importance and potential of AI for drug discovery, this lack of research is surprising. This study aimed to close this research gap by conducting a bibliometric analysis to identify all relevant studies and to analyze interrelationships among algorithms, institutions, countries, and funding sponsors. For this purpose, a sample of 3884 articles was examined bibliometrically, including studies from 1991 to 2022. We utilized various qualitative and quantitative methods, such as performance analysis, science mapping, and thematic analysis. Based on these findings, we furthermore developed a research agenda that aims to serve as a foundation for future researchers.

Artificial intelligence-driven phenotyping of zebrafish psychoactive drug responses

Zebrafish (Danio rerio) are rapidly emerging in biomedicine as promising tools for disease modelling and drug discovery. The use of zebrafish for neuroscience research is also growing rapidly, necessitating novel reliable and unbiased methods of neurophenotypic data collection and analyses. Here, we applied the artificial intelligence (AI) neural network-based algorithms to a large dataset of adult zebrafish locomotor tracks collected previously in a series of in vivo experiments with multiple established psychotropic drugs. We first trained AI to recognize various drugs from a wide range of psychotropic agents tested, and then confirmed prediction accuracy of trained AI by comparing several agents with known similar behavioral and pharmacological profiles. Presenting a framework for innovative neurophenotyping, this proof-of-concept study aims to improve AI-driven movement pattern classification in zebrafish, thereby fostering drug discovery and development utilizing this key model organism.

Encoder-Decoder Full Residual Deep Networks for Robust Regression and Spatiotemporal Estimation

Although increasing hidden layers can improve the ability of a neural network in modeling complex nonlinear relationships, deep layers may result in degradation of accuracy due to the problem of vanishing gradient. Accuracy degradation limits the applications of deep neural networks to predict continuous variables with a small sample size and/or weak or little invariance to translations. Inspired by residual convolutional neural network in computer vision, we developed an encoder-decoder full residual deep network to robustly regress and predict complex spatiotemporal variables. We embedded full shortcuts from each encoding layer to its corresponding decoding layer in a systematic encoder-decoder architecture for efficient residual mapping and error signal propagation. We demonstrated, theoretically and experimentally, that the proposed network structure with full residual connections can successfully boost the backpropagation of signals and improve learning outcomes. This novel method has been extensively evaluated and compared with four commonly used methods (i.e., plain neural network, cascaded residual autoencoder, generalized additive model, and XGBoost) across different testing cases for continuous variable predictions. For model evaluation, we focused on spatiotemporal imputation of satellite aerosol optical depth with massive nonrandomness missingness and spatiotemporal estimation of atmospheric fine particulate matter [Formula: see text] (PM2.5). Compared with the other approaches, our method achieved the state-of-the-art accuracy, had less bias in predicting extreme values, and generated more realistic spatial surfaces. This encoder-decoder full residual deep network can be an efficient and powerful tool in a variety of applications that involve complex nonlinear relationships of continuous variables, varying sample sizes, and spatiotemporal data with weak or little invariance to translation.

CNN-Based Volume Flow Rate Prediction of Oil-Gas-Water Three-Phase Intermittent Flow from Multiple Sensors

In this paper, we propose a deep-learning-based method using a convolutional neural network (CNN) to predict the volume flow rates of individual phases in the oil-gas-water three-phase intermittent flow simultaneously by analyzing the measurement data from multiple sensors, including a temperature sensor, a pressure sensor, a Venturi tube and a microwave sensor. To build datasets, a series of experiments for the oil-gas-water three-phase intermittent flow in a horizontal pipe, in which gas volume fraction and water-in-liquid ratio ranges are 23.77-94.45% and 14.95-86.97%, respectively, and gas flow superficial velocity and liquid flow superficial velocity ranges are 0.66-5.23 and 0.27-2.14 m/s, respectively, have been carried out on a test loop pipeline. The preliminary results indicate that the model can provide relative prediction errors on the testing-1 dataset for the volume flow rates of oil-phase, gas-phase and water-phase within ±10% with 94.49%, 92.56% and 95.71% confidence levels, respectively. Additionally, the prediction results on the testing-2 dataset also demonstrate the generalization ability of the model. The consuming time of a prediction with one sample is 0.43 s on an Intel Xeon CPU E5-2678 v3, and 0.01 s on an NVIDIA GeForce GTX 1080 Ti GPU. Hence, the proposed CNN-based prediction model, which can fulfill the real-time application requirements in the petroleum industry, reveals the potential of using deep learning to obtain accurate results in the multiphase flow measurement field.

DEVELOPMENT OF NEURAL THERMAL SCATTERING (NeTS) MODULES FOR REACTOR MULTI-PHYSICS SIMULATIONS

Modern multi-physics codes, often employed in the simulation and development of thermal nuclear systems, depend heavily on thermal neutron interaction data to determine the space-time distribution of fission events. Therefore, the computationally expensive analysis of such systems motivates the advancement of thermal scattering law (TSL) data delivery methods. Despite considerable improvements on past strategies, current implementations are limited by trade-offs between speed, accuracy, and memory allocation. Furthermore, many of these implementations are not easily adaptable to additional input parameters (e.g., temperature), relying instead on various interpolation schemes. In this work, a novel approach to this problem is demonstrated with a neural network trained on beryllium oxide thermal scattering data generated by the FLASSH nuclear data code of the Low Energy Interaction Physics (LEIP) group at North Carolina State University. Using open-source deep learning libraries, this approach maps a unique functional form to the S(α,β,T) probability distribution function, providing a continuous representation of the TSL across the input phase space. For a given material, the result is a highly accurate, neural thermal scattering (NeTS) module that enables rapid sampling and execution with minimal memory requirements. Moreover, extension of the NeTS phase space to other parameters of interest (e.g., pressure, radiation damage) is highly possible. Consequently, NeTS modules for different materials under various conditions can be stored together in material “lockers” and accessed on-the-fly to generate problem specific cross-sections.

Will We Ever Have Conscious Machines?

The question of whether artificial beings or machines could become self-aware or conscious has been a philosophical question for centuries. The main problem is that self-awareness cannot be observed from an outside perspective and the distinction of being really self-aware or merely a clever imitation cannot be answered without access to knowledge about the mechanism's inner workings. We investigate common machine learning approaches with respect to their potential ability to become self-aware. We realize that many important algorithmic steps toward machines with a core consciousness have already been taken.

Study on Representation Invariances of CNNs and Human Visual Information Processing Based on Data Augmentation

Representation invariance plays a significant role in the performance of deep convolutional neural networks (CNNs) and human visual information processing in various complicated image-based tasks. However, there has been abounding confusion concerning the representation invariance mechanisms of the two sophisticated systems. To investigate their relationship under common conditions, we proposed a representation invariance analysis approach based on data augmentation technology. Firstly, the original image library was expanded by data augmentation. The representation invariances of CNNs and the ventral visual stream were then studied by comparing the similarities of the corresponding layer features of CNNs and the prediction performance of visual encoding models based on functional magnetic resonance imaging (fMRI) before and after data augmentation. Our experimental results suggest that the architecture of CNNs, combinations of convolutional and fully-connected layers, developed representation invariance of CNNs. Remarkably, we found representation invariance belongs to all successive stages of the ventral visual stream. Hence, the internal correlation between CNNs and the human visual system in representation invariance was revealed. Our study promotes the advancement of invariant representation of computer vision and deeper comprehension of the representation invariance mechanism of human visual information processing.

Seven Properties of Self-Organization in the Human Brain

The principle of self-organization has acquired a fundamental significance in the newly emerging field of computational philosophy. Self-organizing systems have been described in various domains in science and philosophy including physics, neuroscience, biology and medicine, ecology, and sociology. While system architecture and their general purpose may depend on domain-specific concepts and definitions, there are (at least) seven key properties of self-organization clearly identified in brain systems: (1) modular connectivity, (2) unsupervised learning, (3) adaptive ability, (4) functional resiliency, (5) functional plasticity, (6) from-local-to-global functional organization, and (7) dynamic system growth. These are defined here in the light of insight from neurobiology, cognitive neuroscience and Adaptive Resonance Theory (ART), and physics to show that self-organization achieves stability and functional plasticity while minimizing structural system complexity. A specific example informed by empirical research is discussed to illustrate how modularity, adaptive learning, and dynamic network growth enable stable yet plastic somatosensory representation for human grip force control. Implications for the design of “strong” artificial intelligence in robotics are brought forward.

Precise Modeling of the Protective Effects of Quercetin against Mycotoxin via System Identification with Neural Networks

Cell cytotoxicity assays, such as cell viability and lactate dehydrogenase (LDH) activity assays, play an important role in toxicological studies of pharmaceutical compounds. However, precise modeling for cytotoxicity studies is essential for successful drug discovery. The aim of our study was to develop a computational modeling that is capable of performing precise prediction, processing, and data representation of cell cytotoxicity. For this, we investigated protective effect of quercetin against various mycotoxins (MTXs), including citrinin (CTN), patulin (PAT), and zearalenol (ZEAR) in four different human cancer cell lines (HeLa, PC-3, Hep G2, and SK-N-MC) in vitro. In addition, the protective effect of quercetin (QCT) against various MTXs was verified via modeling of their nonlinear protective functions using artificial neural networks. The protective model of QCT is built precisely via learning of sparsely measured experimental data by the artificial neural networks (ANNs). The neuromodel revealed that QCT pretreatment at doses of 7.5 to 20 μg/mL significantly attenuated MTX-induced alteration of the cell viability and the LDH activity on HeLa, PC-3, Hep G2, and SK-N-MC cell lines. It has shown that the neuromodel can be used to predict the protective effect of QCT against MTX-induced cytotoxicity for the measurement of percentage (%) of inhibition, cell viability, and LDH activity of MTXs.

Comparison of Machine Learning Regression Algorithms for Cotton Leaf Area Index Retrieval Using Sentinel-2 Spectral Bands

Leaf area index (LAI) is a crucial crop biophysical parameter that has been widely used in a variety of fields. Five state-of-the-art machine learning regression algorithms (MLRAs), namely, artificial neural network (ANN), support vector regression (SVR), Gaussian process regression (GPR), random forest (RF) and gradient boosting regression tree (GBRT), have been used in the retrieval of cotton LAI with Sentinel-2 spectral bands. The performances of the five machine learning models are compared for better applications of MLRAs in remote sensing, since challenging problems remain in the selection of MLRAs for crop LAI retrieval, as well as the decision as to the optimal number for the training sample size and spectral bands to different MLRAs. A comprehensive evaluation was employed with respect to model accuracy, computational efficiency, sensitivity to training sample size and sensitivity to spectral bands. We conducted the comparison of five MLRAs in an agricultural area of Northwest China over three cotton seasons with the corresponding field campaigns for modeling and validation. Results show that the GBRT model outperforms the other models with respect to model accuracy in average ( R 2 ¯ = 0.854, R M S E ¯ = 0.674 and M A E ¯ = 0.456). SVR achieves the best performance in computational efficiency, which means it is fast to train, and to validate that it has great potentials to deliver near-real-time operational products for crop management. As for sensitivity to training sample size, GBRT behaves as the most robust model, and provides the best model accuracy on the average among the variations of training sample size, compared with other models ( R 2 ¯ = 0.884, R M S E ¯ = 0.615 and M A E ¯ = 0.452). Spectral bands sensitivity analysis with dCor (distance correlation), combined with the backward elimination approach, indicates that SVR, GPR and RF provide relatively robust performance to the spectral bands, while ANN outperforms the other models in terms of model accuracy on the average among the reduction of spectral bands ( R 2 ¯ = 0.881, R M S E ¯ = 0.625 and M A E ¯ = 0.480). A comprehensive evaluation indicates that GBRT is an appealing alternative for cotton LAI retrieval, except for its computational efficiency. Despite the different performance of the ML models, all models exhibited considerable potential for cotton LAI retrieval, which could offer accurate crop parameters information timely and accurately for crop fields management and agricultural production decisions.

Deep Neural Networks as Scientific Models

Artificial deep neural networks (DNNs) initially inspired by the brain enable computers to solve cognitive tasks at which humans excel. In the absence of explanations for such cognitive phenomena, in turn cognitive scientists have started using DNNs as models to investigate biological cognition and its neural basis, creating heated debate. Here, we reflect on the case from the perspective of philosophy of science. After putting DNNs as scientific models into context, we discuss how DNNs can fruitfully contribute to cognitive science. We claim that beyond their power to provide predictions and explanations of cognitive phenomena, DNNs have the potential to contribute to an often overlooked but ubiquitous and fundamental use of scientific models: exploration.

Enhancing the Utility of Preclinical Research in Neuropsychiatry Drug Development

Most large pharmaceutical companies have downscaled or closed their clinical neuroscience research programs in response to the low clinical success rate for drugs that showed tremendous promise in animal experiments intended to model psychiatric pathophysiology. These failures have raised serious concerns about the role of preclinical research in the identification and evaluation of new pharmacotherapies for psychiatry. In the absence of a comprehensive understanding of the neurobiology of psychiatric disorders, the task of developing "animal models" seems elusive. The purpose of this review is to highlight emerging strategies to enhance the utility of preclinical research in the drug development process. We address this issue by reviewing how advances in neuroscience, coupled with new conceptual approaches, have recently revolutionized the way we can diagnose and treat common psychiatric conditions. We discuss the implications of these new tools for modeling psychiatric conditions in animals and advocate for the use of systematic reviews of preclinical work as a prerequisite for conducting psychiatric clinical trials. We believe that work in animals is essential for elucidating human psychopathology and that improving the predictive validity of animal models is necessary for developing more effective interventions for mental illness.