Bound to each node is a vector of latent attributes that intuitively represent object properties such as surface shape and texture.

The rise of deep learning challenges the longstanding scientific ideal of insight - the human capacity to understand phenomena by uncovering underlying mechanisms. In many modern applications, accurate predictions no longer require interpretable models, prompting debate about whether explainability is a realistic or even meaningful goal. From our perspective in physics, we examine this tension through a concrete case study: a physics-informed neural network (PINN) trained on a rarefied gas dynamics problem governed by the Boltzmann equation. Despite the system's clear structure and well-understood governing laws, the trained network's weights resemble Gaussian-distributed random matrices, with no evident trace of the physical principles involved. This suggests that deep learning and traditional simulation may follow distinct cognitive paths to the same outcome - one grounded in mechanistic insight, the other in statistical interpolation. Our findings raise critical questions about the limits of explainable AI and whether interpretability can - or should-remain a universal standard in artificial reasoning.

In recent years, physics-informed machine learning has gained significant attention for its ability to enhance statistical performance and sample efficiency by integrating physical structures into machine learning models. These structures, such as differential equations, conservation laws, and symmetries, serve as inductive biases that can improve the generalization capacity of the hybrid model. However, the mechanisms by which these physical structures enhance generalization capacity are not fully understood, limiting the ability to guarantee the performance of the models. In this study, we show that the generalization performance of linear regressors incorporating differential equation structures is determined by the dimension of the associated affine variety, rather than the number of parameters. This finding enables a unified analysis of various equations, including nonlinear ones. We introduce a method to approximate the dimension of the affine variety and provide experimental evidence to validate our theoretical insights. In recent years, physics-informed machine learning (PIML) has garnered significant attention (Rai & Sahu, 2020; Karniadakis et al., 2021; Cuomo et al., 2022; Hao et al., 2022). PIML is a hybrid approach that integrates physical knowledge into machine learning models for tasks involving physical phenomena.

Pre-trained on extensive text and image corpora, current Multi-Modal Large Language Models (MLLM) have shown strong capabilities in general visual reasoning tasks. However, their performance is still lacking in physical domains that require understanding diagrams with complex physical structures and quantitative analysis based on multi-modal information. To address this, we develop a new framework, named Multi-Modal Scientific Reasoning with Physics Perception and Simulation (MAPS) based on an MLLM. MAPS decomposes expert-level multi-modal reasoning task into physical diagram understanding via a Physical Perception Model (PPM) and reasoning with physical knowledge via a simulator. The PPM module is obtained by fine-tuning a visual language model using carefully designed synthetic data with paired physical diagrams and corresponding simulation language descriptions. At the inference stage, MAPS integrates the simulation language description of the input diagram provided by PPM and results obtained through a Chain-of-Simulation process with MLLM to derive the underlying rationale and the final answer. Validated using our collected college-level circuit analysis problems, MAPS significantly improves reasoning accuracy of MLLM and outperforms all existing models. The results confirm MAPS offers a promising direction for enhancing multi-modal scientific reasoning ability of MLLMs. We will release our code, model and dataset used for our experiments upon publishing of this paper.

Large Language Models (LLMs) went from non-existent to ubiquitous in the machine learning discourse within a few years. Due to the fast pace of the field, it is difficult to identify the remaining challenges and already fruitful application areas. In this paper, we aim to establish a systematic set of open problems and application successes so that ML researchers can comprehend the field's current state more quickly and become productive.

Nineteen protesters were arrested at the Kentucky Capitol on Wednesday amid a protest against a measure that would ban certain gender care for minors. Documents obtained by Fox News Digital show that University of Texas Southwestern medical students are being taught that gender is independent of physical structure. Fox News Digital obtained the documents via a FOIA request from Do No Harm, a national association of medical professionals that combats "woke" activism in the healthcare system. According to the University of Texas Southwestern Medical Center's Human Structure curriculum, they "explicitly acknowledge the differentiation between the terms sex and gender." RACHEL LEVINE SAYS CHANGING KIDS' GENDERS WILL SOON BE FULLY EMBRACED: 'WHEELS WILL TURN ON THIS' "The latter is a psychological, social, and cultural construct, including self-identification. Gender is independent of physical structure, chromosomes, or genes," curriculum materials read.

How is it possible to harness high volumes of data on a planetary scale to discover spatial and temporal patterns that escape human perception? The convergence of technologies such as LIDAR and machine learning is allowing for the creation of macroscopes, which have many applications in monitoring and risk analysis for enterprises and governments. Microscopes have been around for centuries, and they are tools that allow individuals to visualize and research phenomena that are too small to be perceived by the human eye. Macroscopes can be thought of as carrying out the opposite function; they are systems that are designed to uncover spatial and temporal patterns that are too large or slow to be perceived by humans. In order to function, they require both the ability to gather planetary-scale information over specified periods of time, as well as the compute technologies that can deal with such data and provide interactive visualization.

This article is part of "the philosophy of artificial intelligence," a series of posts that explore the ethical, moral, and social implications of AI today and in the future. For decades, scientists have tried to create computational imitations of the brain. And for decades, the holy grail of artificial general intelligence, computers that can think and act like humans, has continued to elude scientists and researchers. Why do we continue to replicate some aspects of intelligence but fail to generate systems that can generalize their skills like humans and animals? One computer scientist who has been working on AI for three decades believes that to get past the hurdles of narrow AI, we must look at intelligence from a different and more fundamental perspective.

All the sessions from Transform 2021 are available on-demand now. This article is part of "the philosophy of artificial intelligence," a series of posts that explore the ethical, moral, and social implications of AI today and in the future For decades, scientists have tried to create computational imitations of the brain. And for decades, the holy grail of artificial general intelligence, computers that can think and act like humans, has continued to elude scientists and researchers. Why do we continue to replicate some aspects of intelligence but fail to generate systems that can generalize their skills like humans and animals? One computer scientist who has been working on AI for three decades believes that to get past the hurdles of narrow AI, we must look at intelligence from a different and more fundamental perspective.

This article is part of "the philosophy of artificial intelligence," a series of posts that explore the ethical, moral, and social implications of AI today and in the future For decades, scientists have tried to create computational imitations of the brain. And for decades, the holy grail of artificial general intelligence, computers that can think and act like humans, has continued to elude scientists and researchers. Why do we continue to replicate some aspects of intelligence but fail to generate systems that can generalize their skills like humans and animals? One computer scientist who has been working on AI for three decades believes that to get past the hurdles of narrow AI, we must look at intelligence from a different and more fundamental perspective. In a paper that was presented at the Brain-Inspired Cognitive Architectures for Artificial Intelligence (BICA*AI), Sathyanaraya Raghavachary, Associate Professor of Computer Science at the University of Southern California, discusses "considered response," a theory that can generalize to all forms of intelligent life that have evolved and thrived on our planet.