Liangming Pan
YOU?
Author Swipe
View article: From Atomic to Composite: Reinforcement Learning Enables Generalization in Complementary Reasoning
From Atomic to Composite: Reinforcement Learning Enables Generalization in Complementary Reasoning Open
The mechanism by which RL contributes to reasoning capabilities-whether it incentivizes the synthesis of new skills or merely amplifies existing behaviors-remains a subject of intense debate. In this work, we investigate this question thro…
View article: MuSLR: Multimodal Symbolic Logical Reasoning
MuSLR: Multimodal Symbolic Logical Reasoning Open
Multimodal symbolic logical reasoning, which aims to deduce new facts from multimodal input via formal logic, is critical in high-stakes applications such as autonomous driving and medical diagnosis, as its rigorous, deterministic reasonin…
View article: How Is LLM Reasoning Distracted by Irrelevant Context? An Analysis Using a Controlled Benchmark
How Is LLM Reasoning Distracted by Irrelevant Context? An Analysis Using a Controlled Benchmark Open
We introduce Grade School Math with Distracting Context (GSM-DC), a synthetic benchmark to evaluate Large Language Models' (LLMs) reasoning robustness against systematically controlled irrelevant context (IC). GSM-DC constructs symbolic re…
View article: Numerical investigation on mixed convection in narrow rectangular channel under inclination condition
Numerical investigation on mixed convection in narrow rectangular channel under inclination condition Open
Narrow rectangular channels have been widely adopted in small modular reactors (SMRs), which show particular promise for maintaining stable driving forces under marine operating conditions. The flow and heat transfer characteristics of the…
View article: Combating Multimodal LLM Hallucination via Bottom-Up Holistic Reasoning
Combating Multimodal LLM Hallucination via Bottom-Up Holistic Reasoning Open
Recent advancements in multimodal large language models (MLLMs) have shown unprecedented capabilities in advancing various vision-language tasks. However, MLLMs face significant challenges with hallucinations, and misleading outputs that d…
View article: ConciseRL: Conciseness-Guided Reinforcement Learning for Efficient Reasoning Models
ConciseRL: Conciseness-Guided Reinforcement Learning for Efficient Reasoning Models Open
View article: Phase Field Model for Irradiation-Induced Crack Propagation in Outer Coating Layers of Triso Particle Fuel
Phase Field Model for Irradiation-Induced Crack Propagation in Outer Coating Layers of Triso Particle Fuel Open
View article: How Is LLM Reasoning Distracted by Irrelevant Context? An Analysis Using a Controlled Benchmark
How Is LLM Reasoning Distracted by Irrelevant Context? An Analysis Using a Controlled Benchmark Open
View article: Lead-Cooled Fast Reactor Core Outlet Temperature Oscillation Characteristics and Structural Thermal Fatigue Analysis
Lead-Cooled Fast Reactor Core Outlet Temperature Oscillation Characteristics and Structural Thermal Fatigue Analysis Open
View article: Numerical Study on the Flow Field Characteristics and Efficiency Losses in Nuclear Power Turbines Based on the Non-Equilibrium Condensation Model
Numerical Study on the Flow Field Characteristics and Efficiency Losses in Nuclear Power Turbines Based on the Non-Equilibrium Condensation Model Open
View article: Design and Optimization of He-Xe Brayton Cycles System for Mw-Level Space Nuclear Reactor Application
Design and Optimization of He-Xe Brayton Cycles System for Mw-Level Space Nuclear Reactor Application Open
View article: Experimental and Numerical Study Of Nitrogen Expansion and Condensation Characteristics in Laval Nozzles
Experimental and Numerical Study Of Nitrogen Expansion and Condensation Characteristics in Laval Nozzles Open
View article: Gödel Agent: A Self-Referential Agent Framework for Recursively Self-Improvement
Gödel Agent: A Self-Referential Agent Framework for Recursively Self-Improvement Open
View article: Aristotle: Mastering Logical Reasoning with A Logic-Complete Decompose-Search-Resolve Framework
Aristotle: Mastering Logical Reasoning with A Logic-Complete Decompose-Search-Resolve Framework Open
View article: Effect of Superheat on Non-Equilibrium Condensation in Nuclear Steam Turbines
Effect of Superheat on Non-Equilibrium Condensation in Nuclear Steam Turbines Open
View article: Numerical study on the flow field characteristics and efficiency losses in nuclear power turbines based on the non-equilibrium condensation model
Numerical study on the flow field characteristics and efficiency losses in nuclear power turbines based on the non-equilibrium condensation model Open
View article: AntiLeakBench: Preventing Data Contamination by Automatically Constructing Benchmarks with Updated Real-World Knowledge
AntiLeakBench: Preventing Data Contamination by Automatically Constructing Benchmarks with Updated Real-World Knowledge Open
View article: Investigating the Transferability of Code Repair for Low-Resource Programming Languages
Investigating the Transferability of Code Repair for Low-Resource Programming Languages Open
View article: Design and Optimization of He-Xe Brayton Cycles System for Mw-Level Space Nuclear Reactor Application
Design and Optimization of He-Xe Brayton Cycles System for Mw-Level Space Nuclear Reactor Application Open
View article: RuleArena: A Benchmark for Rule-Guided Reasoning with LLMs in Real-World Scenarios
RuleArena: A Benchmark for Rule-Guided Reasoning with LLMs in Real-World Scenarios Open
View article: Experimental Investigation on Heat Transfer Characteristics of Rotating Heat Pipe
Experimental Investigation on Heat Transfer Characteristics of Rotating Heat Pipe Open
View article: TART: An Open-Source Tool-Augmented Framework for Explainable Table-based Reasoning
TART: An Open-Source Tool-Augmented Framework for Explainable Table-based Reasoning Open
View article: Aristotle: Mastering Logical Reasoning with A Logic-Complete Decompose-Search-Resolve Framework
Aristotle: Mastering Logical Reasoning with A Logic-Complete Decompose-Search-Resolve Framework Open
In the context of large language models (LLMs), current advanced reasoning methods have made impressive strides in various reasoning tasks. However, when it comes to logical reasoning tasks, major challenges remain in both efficacy and eff…
View article: Combating Multimodal LLM Hallucination via Bottom-Up Holistic Reasoning
Combating Multimodal LLM Hallucination via Bottom-Up Holistic Reasoning Open
Recent advancements in multimodal large language models (MLLMs) have shown unprecedented capabilities in advancing various vision-language tasks. However, MLLMs face significant challenges with hallucinations, and misleading outputs that d…
View article: RuleArena: A Benchmark for Rule-Guided Reasoning with LLMs in Real-World Scenarios
RuleArena: A Benchmark for Rule-Guided Reasoning with LLMs in Real-World Scenarios Open
This paper introduces RuleArena, a novel and challenging benchmark designed to evaluate the ability of large language models (LLMs) to follow complex, real-world rules in reasoning. Covering three practical domains -- airline baggage fees,…
View article: TART: An Open-Source Tool-Augmented Framework for Explainable Table-based Reasoning
TART: An Open-Source Tool-Augmented Framework for Explainable Table-based Reasoning Open
Current Large Language Models (LLMs) exhibit limited ability to understand table structures and to apply precise numerical reasoning, which is crucial for tasks such as table question answering (TQA) and table-based fact verification (TFV)…
View article: Investigating the Transferability of Code Repair for Low-Resource Programming Languages
Investigating the Transferability of Code Repair for Low-Resource Programming Languages Open
Large language models (LLMs) have shown remarkable performance on code generation tasks. A recent use case is iterative code repair, where an LLM fixes an incorrect program by rationalizing about errors and generating new code. Recent work…
View article: MultiAgent Collaboration Attack: Investigating Adversarial Attacks in Large Language Model Collaborations via Debate
MultiAgent Collaboration Attack: Investigating Adversarial Attacks in Large Language Model Collaborations via Debate Open
Large Language Models (LLMs) have shown exceptional results on current benchmarks when working individually. The advancement in their capabilities, along with a reduction in parameter size and inference times, has facilitated the use of th…
View article: Modeling Dynamic Topics in Chain-Free Fashion by Evolution-Tracking Contrastive Learning and Unassociated Word Exclusion
Modeling Dynamic Topics in Chain-Free Fashion by Evolution-Tracking Contrastive Learning and Unassociated Word Exclusion Open
Dynamic topic models track the evolution of topics in sequential documents, which have derived various applications like trend analysis and opinion mining. However, existing models suffer from repetitive topic and unassociated topic issues…
View article: Faithful Logical Reasoning via Symbolic Chain-of-Thought
Faithful Logical Reasoning via Symbolic Chain-of-Thought Open
While the recent Chain-of-Thought (CoT) technique enhances the reasoning ability of large language models (LLMs) with the theory of mind, it might still struggle in handling logical reasoning that relies much on symbolic expressions and ri…