arXiv subjects

Jiatong Li

Publications and source records attributed to Jiatong Li.

At least 55 records · Page 3Linked to original sources

arXiv2025

Reasoning Is All You Need for Urban Planning AI

AI has proven highly successful at urban planning analysis -- learning patterns from data to predict future conditions. The next frontier is AI-assisted decision-making: agents that recommend sites, allocate resources, and evaluate trade-offs while reasoning transparently about constraints and stakeholder values. Recent breakthroughs in reasoning AI -- CoT prompting, ReAct, and multi-agent collaboration frameworks -- now make this vision achievable. This position paper presents the Agentic Urban Planning AI Framework for reasoning-capable planning agents that integrates three cognitive layers (Perception, Foundation, Reasoning) with six logic components (Analysis, Generation, Verification, Evaluation, Collaboration, Decision) through a multi-agents collaboration framework. We demonstrate why planning decisions require explicit reasoning capabilities that are value-based (applying normative principles), rule-grounded (guaranteeing constraint satisfaction), and explainable (generating transparent justifications) -- requirements that statistical learning alone cannot fulfill. We compare reasoning agents with statistical learning, present a comprehensive architecture with benchmark evaluation metrics, and outline critical research challenges. This framework shows how AI agents can augment human planners by systematically exploring solution spaces, verifying regulatory compliance, and deliberating over trade-offs transparently -- not replacing human judgment but amplifying it with computational reasoning capabilities.

cs.AI↗

arXiv2025

Chem-R: Learning to Reason as a Chemist

Although large language models (LLMs) have significant potential to advance chemical discovery, current LLMs lack core chemical knowledge, produce unreliable reasoning trajectories, and exhibit suboptimal performance across diverse chemical tasks. To address these challenges, we propose Chem-R, a generalizable Chemical Reasoning model designed to emulate the deliberative processes of chemists. Chem-R is trained through a three-phase framework that progressively builds advanced reasoning capabilities, including: 1) Chemical Foundation Training, which establishes core chemical knowledge. 2) Chemical Reasoning Protocol Distillation, incorporating structured, expert-like reasoning traces to guide systematic and reliable problem solving. 3) Multi-task Group Relative Policy Optimization that optimizes the model for balanced performance across diverse molecular- and reaction-level tasks. This structured pipeline enables Chem-R to achieve state-of-the-art performance on comprehensive benchmarks, surpassing leading large language models, including Gemini-2.5-Pro and DeepSeek-R1, by up to 32% on molecular tasks and 48% on reaction tasks. Meanwhile, Chem-R also consistently outperforms the existing chemical foundation models across both molecular and reaction level tasks. These results highlight Chem-R's robust generalization, interpretability, and potential as a foundation for next-generation AI-driven chemical discovery. The code and model are available at https://github.com/davidweidawang/Chem-R.

cs.CE↗

arXiv2025

Visual Instruction Bottleneck Tuning

Despite widespread adoption, multimodal large language models (MLLMs) suffer performance degradation when encountering unfamiliar queries under distribution shifts. Existing methods to improve MLLM generalization typically require either more instruction data or larger advanced model architectures, both of which incur non-trivial human labor or computational costs. In this work, we take an alternative approach to enhance the generalization and robustness of MLLMs under distribution shifts, from a representation learning perspective. Inspired by information bottleneck (IB) principle, we derive a variational lower bound of the IB for MLLMs and devise a practical implementation, Visual Instruction Bottleneck Tuning (Vittle). We then provide a theoretical justification of Vittle by revealing its connection to an information-theoretic robustness metric of MLLM. Empirical validation of multiple MLLMs on open-ended and closed-form question answering and object hallucination detection tasks over 45 datasets, including 30 shift scenarios, demonstrates that Vittle consistently improves the MLLM's robustness under shifts by pursuing the learning of a minimal sufficient representation.

cs.AI↗

arXiv2025

Learning Hanzi Character Through VR-Based Mortise-Tenon

This paper introduces a novel VR-based system that redefines the acquisition of Hanzi character literacy by integrating traditional mortise-tenon joinery principles (HVRMT).Addressing the challenge of abstract character memorization in digital learning,our system deconstructs Hanzi components into interactive "structural radicals"akin to wooden joint modules.Leveraging PICO's 6DoF spatial tracking and LLM's morphological analysis,learners assemble stroke sequences with haptic feedback simulating wood-to-wood friction.Our system also supports multiplayer online experiences, enhancing engagement and memory retention while preserving intangible cultural heritage. This innovative approach not only enhances engagement and memory retention but also reconstructs the craft wisdom embedded in Chinese writing systems, offering new pathways for preserving intangible cultural heritage in digital ecosystems.For the demo,please refer to this link{https://youtu.be/oUwfFTRpFyo}.

cs.HC↗

arXiv2025

Uniform Inference in High-Dimensional Threshold Regression Models

We develop a uniform inference theory for high-dimensional slope parameters in threshold regression models, allowing for either cross-sectional or time series data. We first establish oracle inequalities for prediction errors, and L1 estimation errors for the Lasso estimator of the slope parameters and the threshold parameter, accommodating heteroskedastic non-subgaussian error terms and non-subgaussian covariates. Next, we derive the asymptotic distribution of tests involving an increasing number of slope parameters by debiasing (or desparsifying) the Lasso estimator in cases with no threshold effect and with a fixed threshold effect. We show that the asymptotic distributions in both cases are the same, allowing us to perform uniform inference without specifying whether the model is a linear or threshold regression. Additionally, we extend the theory to accommodate time series data under the near-epoch dependence assumption. Finally, we identify statistically significant factors influencing cross-country economic growth and quantify the effects of military news shocks on US government spending and GDP, while also estimating a data-driven threshold point in both applications.

econ.EM↗

arXiv2025

CMPhysBench: A Benchmark for Evaluating Large Language Models in Condensed Matter Physics

We introduce CMPhysBench, designed to assess the proficiency of Large Language Models (LLMs) in Condensed Matter Physics, as a novel Benchmark. CMPhysBench is composed of more than 520 graduate-level meticulously curated questions covering both representative subfields and foundational theoretical frameworks of condensed matter physics, such as magnetism, superconductivity, strongly correlated systems, etc. To ensure a deep understanding of the problem-solving process,we focus exclusively on calculation problems, requiring LLMs to independently generate comprehensive solutions. Meanwhile, leveraging tree-based representations of expressions, we introduce the Scalable Expression Edit Distance (SEED) score, which provides fine-grained (non-binary) partial credit and yields a more accurate assessment of similarity between prediction and ground-truth. Our results show that even the best models, Grok-4, reach only 36 average SEED score and 28% accuracy on CMPhysBench, underscoring a significant capability gap, especially for this practical and frontier domain relative to traditional physics. The code anddataset are publicly available at https://github.com/CMPhysBench/CMPhysBench.

cs.LG↗

arXiv2025

Summarize-Exemplify-Reflect: Data-driven Insight Distillation Empowers LLMs for Few-shot Tabular Classification

Recent studies show the promise of large language models (LLMs) for few-shot tabular classification but highlight challenges due to the variability in structured data. To address this, we propose distilling data into actionable insights to enable robust and effective classification by LLMs. Drawing inspiration from human learning processes, we introduce InsightTab, an insight distillation framework guided by principles of divide-and-conquer, easy-first, and reflective learning. Our approach integrates rule summarization, strategic exemplification, and insight reflection through deep collaboration between LLMs and data modeling techniques. The obtained insights enable LLMs to better align their general knowledge and capabilities with the particular requirements of specific tabular tasks. We extensively evaluate InsightTab on nine datasets. The results demonstrate consistent improvement over state-of-the-art methods. Ablation studies further validate the principle-guided distillation process, while analyses emphasize InsightTab's effectiveness in leveraging labeled data and managing bias.

cs.LG↗

arXiv2025

Mol-R1: Towards Explicit Long-CoT Reasoning in Molecule Discovery

Large language models (LLMs), especially Explicit Long Chain-of-Thought (CoT) reasoning models like DeepSeek-R1 and QWQ, have demonstrated powerful reasoning capabilities, achieving impressive performance in commonsense reasoning and mathematical inference. Despite their effectiveness, Long-CoT reasoning models are often criticized for their limited ability and low efficiency in knowledge-intensive domains such as molecule discovery. Success in this field requires a precise understanding of domain knowledge, including molecular structures and chemical principles, which is challenging due to the inherent complexity of molecular data and the scarcity of high-quality expert annotations. To bridge this gap, we introduce Mol-R1, a novel framework designed to improve explainability and reasoning performance of R1-like Explicit Long-CoT reasoning LLMs in text-based molecule generation. Our approach begins with a high-quality reasoning dataset curated through Prior Regulation via In-context Distillation (PRID), a dedicated distillation strategy to effectively generate paired reasoning traces guided by prior regulations. Building upon this, we introduce MoIA, Molecular Iterative Adaptation, a sophisticated training strategy that iteratively combines Supervised Fine-tuning (SFT) with Reinforced Policy Optimization (RPO), tailored to boost the reasoning performance of R1-like reasoning models for molecule discovery. Finally, we examine the performance of Mol-R1 in the text-based molecule reasoning generation task, showing superior performance against existing baselines.

cs.CL↗

arXiv2025

A continuous-wave vacuum ultraviolet laser for the nuclear clock

The exceptionally low-energy isomeric transition in $^{229}$Th at around 148.4 nm offers a unique opportunity for coherent nuclear control and the realisation of a nuclear clock. Recent advances, most notably the incorporation of large ensembles of $^{229}$Th nuclei in transparent crystals and the development of pulsed vacuum-ultraviolet (VUV) lasers, have enabled initial laser spectroscopy of this transition. However, the lack of an intense, narrow-linewidth VUV laser has precluded coherent nuclear manipulation. Here we introduce and demonstrate the first continuous-wave laser at 148.4 nm, generated via four-wave mixing (FWM) in cadmium vapor. The source delivers 100 nW of power with a linewidth well below 100 Hz and supports broad wavelength tunability. This represents a five-orders-of-magnitude improvement in linewidth over all previous single-frequency lasers below 190 nm, marking a major advance in laser technology. We develop a spatially resolved homodyne technique to place a stringent upper bound on the phase noise induced by the FWM process and demonstrate sub-hertz linewidth capability. These results eliminate the final technical hurdle to a $^{229}$Th-based nuclear clock, opening new directions in quantum metrology, nuclear quantum optics and precision tests of the Standard Model. More broadly, they establish a widely tunable, ultranarrow-linewidth laser platform for applications across quantum information science, condensed matter physics, and high-resolution VUV spectroscopy.

physics.atom-ph↗

arXiv2025

Generative Cognitive Diagnosis

Cognitive diagnosis (CD) models latent cognitive states of human learners by analyzing their response patterns on diagnostic tests, serving as a crucial machine learning technique for educational assessment and evaluation. Traditional cognitive diagnosis models typically follow a transductive prediction paradigm that optimizes parameters to fit response scores and extract learner abilities. These approaches face significant limitations as they cannot perform instant diagnosis for new learners without computationally expensive retraining and produce diagnostic outputs with limited reliability. In this study, we introduces a novel generative diagnosis paradigm that fundamentally shifts CD from predictive to generative modeling, enabling inductive inference of cognitive states without parameter re-optimization. We propose two simple yet effective instantiations of this paradigm: Generative Item Response Theory (G-IRT) and Generative Neural Cognitive Diagnosis Model (G-NCDM), which achieve excellent performance improvements over traditional methods. The generative approach disentangles cognitive state inference from response prediction through a well-designed generation process that incorporates identifiability and monotonicity conditions. Extensive experiments on real-world datasets demonstrate the effectiveness of our methodology in addressing scalability and reliability challenges, especially $\times 100$ speedup for the diagnosis of new learners. Our framework opens new avenues for cognitive diagnosis applications in artificial intelligence, particularly for intelligent model evaluation and intelligent education systems. The code is available at https://github.com/CSLiJT/Generative-CD.git.

cs.LG↗

arXiv2025

Control-R: Towards controllable test-time scaling

This paper target in addressing the challenges of underthinking and overthinking in long chain-of-thought (CoT) reasoning for Large Reasoning Models (LRMs) by introducing Reasoning Control Fields (RCF)--a novel test-time approach that injects structured control signals to guide reasoning from a tree search perspective. RCF enables models to adjust reasoning effort according to given control conditions when solving complex tasks. Additionally, we present the Control-R-4K dataset, which consists of challenging problems annotated with detailed reasoning processes and corresponding control fields. To further enhance reasoning control, we propose a Conditional Distillation Finetuning (CDF) method, which trains model--particularly Control-R-32B--to effectively adjust reasoning effort during test time. Experimental results on benchmarks such as AIME2024 and MATH500 demonstrate that our approach achieves state-of-the-art performance at the 32B scale while enabling a controllable Long CoT reasoning process (L-CoT). Overall, this work introduces an effective paradigm for controllable test-time scaling reasoning.

cs.AI↗

arXiv2025

am-ELO: A Stable Framework for Arena-based LLM Evaluation

Arena-based evaluation is a fundamental yet significant evaluation paradigm for modern AI models, especially large language models (LLMs). Existing framework based on ELO rating system suffers from the inevitable instability problem due to ranking inconsistency and the lack of attention to the varying abilities of annotators. In this paper, we introduce a novel stable arena framework to address these issues by enhancing the ELO Rating System. Specifically, we replace the iterative update method with a Maximum Likelihood Estimation (MLE) approach, m-ELO, and provide theoretical proof of the consistency and stability of the MLE approach for model ranking. Additionally, we proposed the am-ELO, which modify the Elo Rating's probability function to incorporate annotator abilities, enabling the simultaneous estimation of model scores and annotator reliability. Experiments demonstrate that this method ensures stability, proving that this framework offers a more robust, accurate, and stable evaluation method for LLMs.

cs.AI↗

arXiv2025

Benchmarking Poisoning Attacks against Retrieval-Augmented Generation

Retrieval-Augmented Generation (RAG) has proven effective in mitigating hallucinations in large language models by incorporating external knowledge during inference. However, this integration introduces new security vulnerabilities, particularly to poisoning attacks. Although prior work has explored various poisoning strategies, a thorough assessment of their practical threat to RAG systems remains missing. To address this gap, we propose the first comprehensive benchmark framework for evaluating poisoning attacks on RAG. Our benchmark covers 5 standard question answering (QA) datasets and 10 expanded variants, along with 13 poisoning attack methods and 7 defense mechanisms, representing a broad spectrum of existing techniques. Using this benchmark, we conduct a comprehensive evaluation of all included attacks and defenses across the full dataset spectrum. Our findings show that while existing attacks perform well on standard QA datasets, their effectiveness drops significantly on the expanded versions. Moreover, our results demonstrate that various advanced RAG architectures, such as sequential, branching, conditional, and loop RAG, as well as multi-turn conversational RAG, multimodal RAG systems, and RAG-based LLM agent systems, remain susceptible to poisoning attacks. Notably, current defense techniques fail to provide robust protection, underscoring the pressing need for more resilient and generalizable defense strategies.

cs.CR↗

arXiv2025

Critic-V: VLM Critics Help Catch VLM Errors in Multimodal Reasoning

Vision-language models (VLMs) have shown remarkable advancements in multimodal reasoning tasks. However, they still often generate inaccurate or irrelevant responses due to issues like hallucinated image understandings or unrefined reasoning paths. To address these challenges, we introduce Critic-V, a novel framework inspired by the Actor-Critic paradigm to boost the reasoning capability of VLMs. This framework decouples the reasoning process and critic process by integrating two independent components: the Reasoner, which generates reasoning paths based on visual and textual inputs, and the Critic, which provides constructive critique to refine these paths. In this approach, the Reasoner generates reasoning responses according to text prompts, which can evolve iteratively as a policy based on feedback from the Critic. This interaction process was theoretically driven by a reinforcement learning framework where the Critic offers natural language critiques instead of scalar rewards, enabling more nuanced feedback to boost the Reasoner's capability on complex reasoning tasks. The Critic model is trained using Direct Preference Optimization (DPO), leveraging a preference dataset of critiques ranked by Rule-based Reward~(RBR) to enhance its critic capabilities. Evaluation results show that the Critic-V framework significantly outperforms existing methods, including GPT-4V, on 5 out of 8 benchmarks, especially regarding reasoning accuracy and efficiency. Combining a dynamic text-based policy for the Reasoner and constructive feedback from the preference-optimized Critic enables a more reliable and context-sensitive multimodal reasoning process. Our approach provides a promising solution to enhance the reliability of VLMs, improving their performance in real-world reasoning-heavy multimodal applications such as autonomous driving and embodied intelligence.

cs.CV↗

arXiv2025

Large Language Models are In-Context Molecule Learners

Large Language Models (LLMs) have demonstrated exceptional performance in biochemical tasks, especially the molecule caption translation task, which aims to bridge the gap between molecules and natural language texts. However, previous methods in adapting LLMs to the molecule-caption translation task required extra domain-specific pre-training stages, suffered weak alignment between molecular and textual spaces, or imposed stringent demands on the scale of LLMs. To resolve the challenges, we propose In-Context Molecule Adaptation (ICMA), as a new paradigm allowing LLMs to learn the molecule-text alignment from context examples via In-Context Molecule Tuning. Specifically, ICMA incorporates the following three stages: Hybrid Context Retrieval, Post-retrieval Re-ranking, and In-context Molecule Tuning. Initially, Hybrid Context Retrieval utilizes BM25 Caption Retrieval and Molecule Graph Retrieval to retrieve similar informative context examples. Additionally, Post-retrieval Re-ranking is composed of Sequence Reversal and Random Walk selection to further improve the quality of retrieval results. Finally, In-Context Molecule Tuning unlocks the in-context learning and reasoning capability of LLMs with the retrieved examples and adapts the parameters of LLMs for better alignment between molecules and texts. Experimental results demonstrate that ICMA can empower LLMs to achieve state-of-the-art or comparable performance without extra training corpora and intricate structures, showing that LLMs are inherently in-context molecule learners.

cs.CL↗

arXiv2024

An Experimental Study Of Netflix Use and the Effects of Autoplay on Watching Behaviors

Prior work on dark patterns, or manipulative online interfaces, suggests they have potentially detrimental effects on user autonomy. Dark pattern features, like those designed for attention capture, can potentially extend platform sessions beyond that users would have otherwise intended. Existing research, however, has not formally measured the quantitative effects of these features on user engagement in subscription video-on-demand platforms (SVODs). In this work, we conducted an experimental study with 76 Netflix users in the US to analyze the impact of a specific attention capture feature, autoplay, on key viewing metrics. We found that disabling autoplay on Netflix significantly reduced key content consumption aggregates, including average daily watching and average session length, partly filling the evidentiary gap regarding the empirical effects of dark pattern interfaces. We paired the experimental analysis with users' perceptions of autoplay and their viewing behaviors, finding that participants were split on whether the effects of autoplay outweigh its benefits, albeit without knowledge of the study findings. Our findings strengthen the broader argument that manipulative interface designs can and do affect users in potentially damaging ways, highlighting the continued need for considering user well-being and varied preferences in interface design.

cs.HC↗

arXiv2024

LLaMA-Berry: Pairwise Optimization for O1-like Olympiad-Level Mathematical Reasoning

This paper presents an advanced mathematical problem-solving framework, LLaMA-Berry, for enhancing the mathematical reasoning ability of Large Language Models (LLMs). The framework combines Monte Carlo Tree Search (MCTS) with iterative Self-Refine to optimize the reasoning path and utilizes a pairwise reward model to evaluate different paths globally. By leveraging the self-critic and rewriting capabilities of LLMs, Self-Refine applied to MCTS (SR-MCTS) overcomes the inefficiencies and limitations of conventional step-wise and greedy search algorithms by fostering a more efficient exploration of solution spaces. Pairwise Preference Reward Model~(PPRM), inspired by Reinforcement Learning from Human Feedback (RLHF), is then used to model pairwise preferences between solutions, utilizing an Enhanced Borda Count (EBC) method to synthesize these preferences into a global ranking score to find better answers. This approach addresses the challenges of scoring variability and non-independent distributions in mathematical reasoning tasks. The framework has been tested on general and advanced benchmarks, showing superior performance in terms of search efficiency and problem-solving capability compared to existing methods like ToT and rStar, particularly in complex Olympiad-level benchmarks, including GPQA, AIME24 and AMC23.

cs.AI↗

arXiv2024

Leveraging LLMs for Hypothetical Deduction in Logical Inference: A Neuro-Symbolic Approach

Large Language Models (LLMs) have exhibited remarkable potential across a wide array of reasoning tasks, including logical reasoning. Although massive efforts have been made to empower the logical reasoning ability of LLMs via external logical symbolic solvers, crucial challenges of the poor generalization ability to questions with different features and inevitable question information loss of symbolic solver-driven approaches remain unresolved. To mitigate these issues, we introduce LINA, a LLM-driven neuro-symbolic approach for faithful logical reasoning. By enabling an LLM to autonomously perform the transition from propositional logic extraction to sophisticated logical reasoning, LINA not only bolsters the resilience of the reasoning process but also eliminates the dependency on external solvers. Additionally, through its adoption of a hypothetical-deductive reasoning paradigm, LINA effectively circumvents the expansive search space challenge that plagues traditional forward reasoning methods. Empirical evaluations demonstrate that LINA substantially outperforms both established propositional logic frameworks and conventional prompting techniques across a spectrum of five logical reasoning tasks. Specifically, LINA achieves an improvement of 24.34% over LINC on the FOLIO dataset, while also surpassing prompting strategies like CoT and CoT-SC by up to 24.02%. Our code is available at https://github.com/wufeiwuwoshihua/nshy.

cs.CL↗