Explainable neural networks that simulate reasoning

UNSEG: unsupervised segmentation of cells and their nuclei in complex tissue samples

Multiplexed imaging technologies have made it possible to interrogate complex tissue microenvironments at sub-cellular resolution within their native spatial context. However, proper quantification of this complexity requires the ability to easily and accurately segment cells into their sub-cellular compartments. Within the supervised learning paradigm, deep learning-based segmentation methods demonstrating human level performance have emerged. However, limited work has been done in developing such generalist methods within the unsupervised context. Here we present an easy-to-use unsupervised segmentation (UNSEG) method that achieves deep learning level performance without requiring any training data via leveraging a Bayesian-like framework, and nucleus and cell membrane markers. We show that UNSEG is internally consistent and better at generalizing to the complexity of tissue morphology than current deep learning methods, allowing it to unambiguously identify the cytoplasmic compartment of a cell, and localize molecules to their correct sub-cellular compartment. We also introduce a perturbed watershed algorithm for stably and automatically segmenting a cluster of cell nuclei into individual nuclei that increases the accuracy of classical watershed. Finally, we demonstrate the efficacy of UNSEG on a high-quality annotated gastrointestinal tissue dataset we have generated, on publicly available datasets, and in a range of practical scenarios.

UNSEG: unsupervised segmentation of cells and their nuclei in complex tissue samples

Multiplexed imaging technologies have made it possible to interrogate complex tumor microenvironments at sub-cellular resolution within their native spatial context. However, proper quantification of this complexity requires the ability to easily and accurately segment cells into their sub-cellular compartments. Within the supervised learning paradigm, deep learning based segmentation methods demonstrating human level performance have emerged. However, limited work has been done in developing such generalist methods within the label-free unsupervised context. Here we present an unsupervised segmentation (UNSEG) method that achieves deep learning level performance without requiring any training data. UNSEG leverages a Bayesian-like framework and the specificity of nucleus and cell membrane markers to construct an a posteriori probability estimate of each pixel belonging to the nucleus, cell membrane, or background. It uses this estimate to segment each cell into its nuclear and cell-membrane compartments. We show that UNSEG is more internally consistent and better at generalizing to the complexity of tissue morphology than current deep learning methods. This allows UNSEG to unambiguously identify the cytoplasmic compartment of a cell, which we employ to demonstrate its use in an exemplar biological scenario. Within the UNSEG framework, we also introduce a new perturbed watershed algorithm capable of stably and automatically segmenting a cluster of cell nuclei into individual cell nuclei that increases the accuracy of classical watershed. Perturbed watershed can also be used as a standalone algorithm that researchers can incorporate within their supervised or unsupervised learning approaches to extend classical watershed, particularly in the multiplexed imaging context. Finally, as part of developing UNSEG, we have generated a high-quality annotated gastrointestinal tissue (GIT) dataset, which we anticipate will be useful for the broader research community. We demonstrate the efficacy of UNSEG on the GIT dataset, publicly available datasets, and on a range of practical scenarios. In these contexts, we also discuss the possibility of bias inherent in quantification of segmentation accuracy based on F 1 score. Segmentation, despite its long antecedents, remains a challenging problem, particularly in the context of tissue samples. UNSEG, an easy-to-use algorithm, provides an unsupervised approach to overcome this bottleneck, and as we discuss, can help improve deep learning based segmentation methods by providing a bridge between unsupervised and supervised learning paradigms.

Physics-Informed Generative Model for Drug-like Molecule Conformers

We present a diffusion-based generative model for conformer generation. Our model is focused on the reproduction of the bonded structure and is constructed from the associated terms traditionally found in classical force fields to ensure a physically relevant representation. Techniques in deep learning are used to infer atom typing and geometric parameters from a training set. Conformer sampling is achieved by taking advantage of recent advancements in diffusion-based generation. By training on large, synthetic data sets of diverse, drug-like molecules optimized with the semiempirical GFN2-xTB method, high accuracy is achieved for bonded parameters, exceeding that of conventional, knowledge-based methods. Results are also compared to experimental structures from the Protein Databank and the Cambridge Structural Database.

Advancements and challenges of digital twins in industry

Digital twins, which are considered an effective approach to realize the fusion between virtual and physical spaces, have attracted a substantial amount of attention in the past decade. With their rapid development in recent years, digital twins have been applied in various fields, particularly in industry. However, there are still some gaps to be filled and some limitations to be addressed. Here we provide a brief overview of digital twin advancements in industry and highlight the main pitfalls to avoid and challenges to overcome, to improve the maturity of digital twins and facilitate large-scale industrial applications in the future.

Automated discovery of algorithms from data

To automate the discovery of new scientific and engineering principles, artificial intelligence must distill explicit rules from experimental data. This has proven difficult because existing methods typically search through the enormous space of possible functions. Here we introduce deep distilling, a machine learning method that does not perform searches but instead learns from data using symbolic essence neural networks and then losslessly condenses the network parameters into a concise algorithm written in computer code. This distilled code, which can contain loops and nested logic, is equivalent to the neural network but is human-comprehensible and orders-of-magnitude more compact. On arithmetic, vision and optimization tasks, the distilled code is capable of out-of-distribution systematic generalization to solve cases orders-of-magnitude larger and more complex than the training data. The distilled algorithms can sometimes outperform human-designed algorithms, demonstrating that deep distilling is able to discover generalizable principles complementary to human expertise.

Is AI essential? Examining the need for deep learning in image-activated sorting of Saccharomyces cerevisiae

Artificial intelligence (AI) has become a focal point across a multitude of societal sectors, with science not being an exception. Particularly in the life sciences, imaging flow cytometry has increasingly integrated AI for automated management and categorization of extensive cell image data. However, the necessity of AI over traditional classification methods when extending imaging flow cytometry to include cell sorting remains uncertain, primarily due to the time constraints between image acquisition and sorting actuation. AI-enabled image-activated cell sorting (IACS) methods remain substantially limited, even as recent advancements in IACS have found success while largely relying on traditional feature gating strategies. Here we assess the necessity of AI for image classification in IACS by contrasting the performance of feature gating, classical machine learning (ML), and deep learning (DL) with convolutional neural networks (CNNs) in the differentiation of Saccharomyces cerevisiae mutant images. We show that classical ML could only yield a 2.8-fold enhancement in target enrichment capability, albeit at the cost of a 13.7-fold increase in processing time. Conversely, a CNN could offer an 11.0-fold improvement in enrichment capability at an 11.5-fold increase in processing time. We further executed IACS on mixed mutant populations and quantified target strain enrichment via downstream DNA sequencing to substantiate the applicability of DL for the proposed study. Our findings validate the feasibility and value of employing DL in IACS for morphology-based genetic screening of S. cerevisiae, encouraging its incorporation in future advancements of similar technologies.

Achieving descriptive accuracy in explanations via argumentation: The case of probabilistic classifiers

The pursuit of trust in and fairness of AI systems in order to enable human-centric goals has been gathering pace of late, often supported by the use of explanations for the outputs of these systems. Several properties of explanations have been highlighted as critical for achieving trustworthy and fair AI systems, but one that has thus far been overlooked is that of descriptive accuracy (DA), i.e., that the explanation contents are in correspondence with the internal working of the explained system. Indeed, the violation of this core property would lead to the paradoxical situation of systems producing explanations which are not suitably related to how the system actually works: clearly this may hinder user trust. Further, if explanations violate DA then they can be deceitful, resulting in an unfair behavior toward the users. Crucial as the DA property appears to be, it has been somehow overlooked in the XAI literature to date. To address this problem, we consider the questions of formalizing DA and of analyzing its satisfaction by explanation methods. We provide formal definitions of naive, structural and dialectical DA, using the family of probabilistic classifiers as the context for our analysis. We evaluate the satisfaction of our given notions of DA by several explanation methods, amounting to two popular feature-attribution methods from the literature, variants thereof and a novel form of explanation that we propose. We conduct experiments with a varied selection of concrete probabilistic classifiers and highlight the importance, with a user study, of our most demanding notion of dialectical DA, which our novel method satisfies by design and others may violate. We thus demonstrate how DA could be a critical component in achieving trustworthy and fair systems, in line with the principles of human-centric AI.

Semantic Interpretation for Convolutional Neural Networks: What Makes a Cat a Cat?

The interpretability of deep neural networks has attracted increasing attention in recent years, and several methods have been created to interpret the "black box" model. Fundamental limitations remain, however, that impede the pace of understanding the networks, especially the extraction of understandable semantic space. In this work, the framework of semantic explainable artificial intelligence (S-XAI) is introduced, which utilizes a sample compression method based on the distinctive row-centered principal component analysis (PCA) that is different from the conventional column-centered PCA to obtain common traits of samples from the convolutional neural network (CNN), and extracts understandable semantic spaces on the basis of discovered semantically sensitive neurons and visualization techniques. Statistical interpretation of the semantic space is also provided, and the concept of semantic probability is proposed. The experimental results demonstrate that S-XAI is effective in providing a semantic interpretation for the CNN, and offers broad usage, including trustworthiness assessment and semantic sample searching.

Medical Professional Enhancement Using Explainable Artificial Intelligence in Fetal Cardiac Ultrasound Screening

Diagnostic support tools based on artificial intelligence (AI) have exhibited high performance in various medical fields. However, their clinical application remains challenging because of the lack of explanatory power in AI decisions (black box problem), making it difficult to build trust with medical professionals. Nevertheless, visualizing the internal representation of deep neural networks will increase explanatory power and improve the confidence of medical professionals in AI decisions. We propose a novel deep learning-based explainable representation "graph chart diagram" to support fetal cardiac ultrasound screening, which has low detection rates of congenital heart diseases due to the difficulty in mastering the technique. Screening performance improves using this representation from 0.966 to 0.975 for experts, 0.829 to 0.890 for fellows, and 0.616 to 0.748 for residents in the arithmetic mean of area under the curve of a receiver operating characteristic curve. This is the first demonstration wherein examiners used deep learning-based explainable representation to improve the performance of fetal cardiac ultrasound screening, highlighting the potential of explainable AI to augment examiner capabilities.

An ensemble approach to the structure-function problem in microbial communities

The metabolic activity of microbial communities plays a primary role in the flow of essential nutrients throughout the biosphere. Molecular genetics has revealed the metabolic pathways that model organisms utilize to generate energy and biomass, but we understand little about how the metabolism of diverse, natural communities emerges from the collective action of its constituents. We propose that quantifying and mapping metabolic fluxes to sequencing measurements of genomic, taxonomic, or transcriptional variation across an ensemble of diverse communities, either in the laboratory or in the wild, can reveal low-dimensional descriptions of community structure that can explain or predict their emergent metabolic activity. We survey the types of communities for which this approach might be best suited, review the analytical techniques available for quantifying metabolite fluxes in communities, and discuss what types of data analysis approaches might be lucrative for learning the structure-function mapping in communities from these data.