R1-VL: Learning to Reason with Multimodal Large Language Models via Step-wise Group Relative Policy Optimization

Recent studies generally enhance MLLMs' reasoning capabilities via supervised fine-tuning on high-quality chain-of-thought reasoning data, which often leads models to merely imitate successful reasoning paths without understanding what the wrong reasoning paths are. In this work, we aim to enhance the MLLMs' reasoning ability beyond passively imitating positive reasoning paths. To this end, we design Step-wise Group Relative Policy Optimization (StepGRPO), a new online reinforcement learning framework that enables MLLMs to self-improve reasoning ability via simple, effective and dense step-wise rewarding. Specifically, StepGRPO introduces two novel rule-based reasoning rewards: Step-wise Reasoning Accuracy Reward (StepRAR) and Step-wise Reasoning Validity Reward (StepRVR). StepRAR rewards the reasoning paths that contain necessary intermediate reasoning steps via a soft key-step matching technique, while StepRAR rewards reasoning paths that follow a well-structured and logically consistent reasoning process through a reasoning completeness and logic evaluation strategy. With the proposed StepGRPO, we introduce R1-VL, a series of MLLMs with outstanding capabilities in step-by-step reasoning. Extensive experiments over 8 benchmarks demonstrate the superiority of our methods.

Benchmarking Reasoning Robustness in Large Language Models

Despite the recent success of large language models (LLMs) in reasoning such as DeepSeek, we for the first time identify a key dilemma in reasoning robustness and generalization: significant performance degradation on novel or incomplete data, suggesting a reliance on memorized patterns rather than systematic reasoning. Our closer examination reveals four key unique limitations underlying this issue:(1) Positional bias--models favor earlier queries in multi-query inputs but answering the wrong one in the latter (e.g., GPT-4o's accuracy drops from 75.8 percent to 72.8 percent); (2) Instruction sensitivity--performance declines by 5.0 to 7.5 percent in the Qwen2.5 Series and by 5.0 percent in DeepSeek-V3 with auxiliary guidance; (3) Numerical fragility--value substitution sharply reduces accuracy (e.g., GPT-4o drops from 97.5 percent to 82.5 percent, GPT-o1-mini drops from 97.5 percent to 92.5 percent); and (4) Memory dependence--models resort to guesswork when missing critical data. These findings further highlight the reliance on heuristic recall over rigorous logical inference, demonstrating challenges in reasoning robustness. To comprehensively investigate these robustness challenges, this paper introduces a novel benchmark, termed as Math-RoB, that exploits hallucinations triggered by missing information to expose reasoning gaps. This is achieved by an instruction-based approach to generate diverse datasets that closely resemble training distributions, facilitating a holistic robustness assessment and advancing the development of more robust reasoning frameworks. Bad character(s) in field Abstract.

JustLogic: A Comprehensive Benchmark for Evaluating Deductive Reasoning in Large Language Models

Logical reasoning is a critical component of Large Language Models (LLMs), and substantial research efforts in recent years have aimed to enhance their deductive reasoning capabilities. However, existing deductive reasoning benchmarks, which are crucial for evaluating and advancing LLMs, are inadequate due to their lack of task complexity, presence of prior knowledge as a confounder, and superficial error analysis. To address these deficiencies, we introduce JustLogic, a synthetically generated deductive reasoning benchmark designed for rigorous evaluation of LLMs. JustLogic is (i) highly complex, capable of generating a diverse range of linguistic patterns, vocabulary, and argument structures; (ii) prior knowledge independent, eliminating the advantage of models possessing prior knowledge and ensuring that only deductive reasoning is used to answer questions; and (iii) capable of in-depth error analysis on the heterogeneous effects of reasoning depth and argument form on model accuracy. Our experimental results on JustLogic reveal that most state-of-the-art (SOTA) LLMs perform significantly worse than the human average, demonstrating substantial room for model improvement. All code and data are available at https://github.com/michaelchen-lab/JustLogic

How to Build the Virtual Cell with Artificial Intelligence: Priorities and Opportunities

The cell is arguably the smallest unit of life and is central to understanding biology. Accurate modeling of cells is important for this understanding as well as for determining the root causes of disease. Recent advances in artificial intelligence (AI), combined with the ability to generate large-scale experimental data, present novel opportunities to model cells. Here we propose a vision of AI-powered Virtual Cells, where robust representations of cells and cellular systems under different conditions are directly learned from growing biological data across measurements and scales. We discuss desired capabilities of AI Virtual Cells, including generating universal representations of biological entities across scales, and facilitating interpretable in silico experiments to predict and understand their behavior using Virtual Instruments. We further address the challenges, opportunities and requirements to realize this vision including data needs, evaluation strategies, and community standards and engagement to ensure biological accuracy and broad utility. We envision a future where AI Virtual Cells help identify new drug targets, predict cellular responses to perturbations, as well as scale hypothesis exploration. With open science collaborations across the biomedical ecosystem that includes academia, philanthropy, and the biopharma and AI industries, a comprehensive predictive understanding of cell mechanisms and interactions is within reach.

Learning System Dynamics without Forgetting

Predicting the trajectories of systems with unknown dynamics (\textit{i.e.} the governing rules) is crucial in various research fields, including physics and biology. This challenge has gathered significant attention from diverse communities. Most existing works focus on learning fixed system dynamics within one single system. However, real-world applications often involve multiple systems with different types of dynamics or evolving systems with non-stationary dynamics (dynamics shifts). When data from those systems are continuously collected and sequentially fed to machine learning models for training, these models tend to be biased toward the most recently learned dynamics, leading to catastrophic forgetting of previously observed/learned system dynamics. To this end, we aim to learn system dynamics via continual learning. Specifically, we present a novel framework of Mode-switching Graph ODE (MS-GODE), which can continually learn varying dynamics and encode the system-specific dynamics into binary masks over the model parameters. During the inference stage, the model can select the most confident mask based on the observational data to identify the system and predict future trajectories accordingly. Empirically, we systematically investigate the task configurations and compare the proposed MS-GODE with state-of-the-art techniques. More importantly, we construct a novel benchmark of biological dynamic systems, featuring diverse systems with disparate dynamics and significantly enriching the research field of machine learning for dynamic systems.

Continual Learning on Graphs: Challenges, Solutions, and Opportunities

Continual learning on graph data has recently attracted paramount attention for its aim to resolve the catastrophic forgetting problem on existing tasks while adapting the sequentially updated model to newly emerged graph tasks. While there have been efforts to summarize progress on continual learning research over Euclidean data, e.g., images and texts, a systematic review of progress in continual learning on graphs, a.k.a, continual graph learning (CGL) or lifelong graph learning, is still demanding. Graph data are far more complex in terms of data structures and application scenarios, making CGL task settings, model designs, and applications extremely challenging. To bridge the gap, we provide a comprehensive review of existing continual graph learning (CGL) algorithms by elucidating the different task settings and categorizing the existing methods based on their characteristics. We compare the CGL methods with traditional continual learning techniques and analyze the applicability of the traditional continual learning techniques to CGL tasks. Additionally, we review the benchmark works that are crucial to CGL research. Finally, we discuss the remaining challenges and propose several future directions. We will maintain an up-to-date GitHub repository featuring a comprehensive list of CGL algorithms, accessible at https://github.com/UConn-DSIS/Survey-of-Continual-Learning-on-Graphs.

Empowering Time Series Analysis with Large Language Models: A Survey

Recently, remarkable progress has been made over large language models (LLMs), demonstrating their unprecedented capability in varieties of natural language tasks. However, completely training a large general-purpose model from the scratch is challenging for time series analysis, due to the large volumes and varieties of time series data, as well as the non-stationarity that leads to concept drift impeding continuous model adaptation and re-training. Recent advances have shown that pre-trained LLMs can be exploited to capture complex dependencies in time series data and facilitate various applications. In this survey, we provide a systematic overview of existing methods that leverage LLMs for time series analysis. Specifically, we first state the challenges and motivations of applying language models in the context of time series as well as brief preliminaries of LLMs. Next, we summarize the general pipeline for LLM-based time series analysis, categorize existing methods into different groups (i.e., direct query, tokenization, prompt design, fine-tune, and model integration), and highlight the key ideas within each group. We also discuss the applications of LLMs for both general and spatial-temporal time series data, tailored to specific domains. Finally, we thoroughly discuss future research opportunities to empower time series analysis with LLMs.

Topology-aware Embedding Memory for Learning on Expanding Graphs

Memory replay based techniques have shown great success for continual learning with incrementally accumulated Euclidean data. Directly applying them to continually expanding graphs, however, leads to the potential memory explosion problem due to the need to buffer representative nodes and their associated topological neighborhood structures. To this end, we systematically analyze the key challenges in the memory explosion problem, and present a general framework, i.e., Parameter Decoupled Graph Neural Networks (PDGNNs) with Topology-aware Embedding Memory (TEM), to tackle this issue. The proposed framework not only reduces the memory space complexity from $\mathcal{O}(nd^L)$ to $\mathcal{O}(n)$~\footnote{$n$: memory budget, $d$: average node degree, $L$: the radius of the GNN receptive field}, but also fully utilizes the topological information for memory replay. Specifically, PDGNNs decouple trainable parameters from the computation ego-subgraph via \textit{Topology-aware Embeddings} (TEs), which compress ego-subgraphs into compact vectors (i.e., TEs) to reduce the memory consumption. Based on this framework, we discover a unique \textit{pseudo-training effect} in continual learning on expanding graphs and this effect motivates us to develop a novel \textit{coverage maximization sampling} strategy that can enhance the performance with a tight memory budget. Thorough empirical studies demonstrate that, by tackling the memory explosion problem and incorporating topological information into memory replay, PDGNNs with TEM significantly outperform state-of-the-art techniques, especially in the challenging class-incremental setting.

CLEVRER-Humans: Describing Physical and Causal Events the Human Way

Building machines that can reason about physical events and their causal relationships is crucial for flexible interaction with the physical world. However, most existing physical and causal reasoning benchmarks are exclusively based on synthetically generated events and synthetic natural language descriptions of causal relationships. This design brings up two issues. First, there is a lack of diversity in both event types and natural language descriptions; second, causal relationships based on manually-defined heuristics are different from human judgments. To address both shortcomings, we present the CLEVRER-Humans benchmark, a video reasoning dataset for causal judgment of physical events with human labels. We employ two techniques to improve data collection efficiency: first, a novel iterative event cloze task to elicit a new representation of events in videos, which we term Causal Event Graphs (CEGs); second, a data augmentation technique based on neural language generative models. We convert the collected CEGs into questions and answers to be consistent with prior work. Finally, we study a collection of baseline approaches for CLEVRER-Humans question-answering, highlighting the great challenges set forth by our benchmark.

Deep Bidirectional Language-Knowledge Graph Pretraining

Pretraining a language model (LM) on text has been shown to help various downstream NLP tasks. Recent works show that a knowledge graph (KG) can complement text data, offering structured background knowledge that provides a useful scaffold for reasoning. However, these works are not pretrained to learn a deep fusion of the two modalities at scale, limiting the potential to acquire fully joint representations of text and KG. Here we propose DRAGON (Deep Bidirectional Language-Knowledge Graph Pretraining), a self-supervised approach to pretraining a deeply joint language-knowledge foundation model from text and KG at scale. Specifically, our model takes pairs of text segments and relevant KG subgraphs as input and bidirectionally fuses information from both modalities. We pretrain this model by unifying two self-supervised reasoning tasks, masked language modeling and KG link prediction. DRAGON outperforms existing LM and LM+KG models on diverse downstream tasks including question answering across general and biomedical domains, with +5% absolute gain on average. In particular, DRAGON achieves notable performance on complex reasoning about language and knowledge (+10% on questions involving long contexts or multi-step reasoning) and low-resource QA (+8% on OBQA and RiddleSense), and new state-of-the-art results on various BioNLP tasks. Our code and trained models are available at https://github.com/michiyasunaga/dragon.