Yunpu Ma
YOU?
Author Swipe
View article: Beyond Magic Words: Sharpness-Aware Prompt Evolving for Robust Large Language Models with TARE
Beyond Magic Words: Sharpness-Aware Prompt Evolving for Robust Large Language Models with TARE Open
The performance of Large Language Models (LLMs) hinges on carefully engineered prompts. However, prevailing prompt optimization methods, ranging from heuristic edits and reinforcement learning to evolutionary search, primarily target point…
View article: First Experience with Real-Time Control Using Simulated VQC-Based Quantum Policies
First Experience with Real-Time Control Using Simulated VQC-Based Quantum Policies Open
This paper investigates the integration of quantum computing into offline reinforcement learning and the deployment of the resulting quantum policy in a real-time control hardware realization of the cart-pole system. Variational Quantum Ci…
View article: SwarmAgentic: Towards Fully Automated Agentic System Generation via Swarm Intelligence
SwarmAgentic: Towards Fully Automated Agentic System Generation via Swarm Intelligence Open
The rapid progress of Large Language Models has advanced agentic systems in decision-making, coordination, and task execution. Yet, existing agentic system generation frameworks lack full autonomy, missing from-scratch agent generation, se…
View article: ImpliRet: Benchmarking the Implicit Fact Retrieval Challenge
ImpliRet: Benchmarking the Implicit Fact Retrieval Challenge Open
Retrieval systems are central to many NLP pipelines, but often rely on surface-level cues such as keyword overlap and lexical semantic similarity. To evaluate retrieval beyond these shallow signals, recent benchmarks introduce reasoning-he…
View article: ASCD: Attention-Steerable Contrastive Decoding for Reducing Hallucination in MLLM
ASCD: Attention-Steerable Contrastive Decoding for Reducing Hallucination in MLLM Open
Multimodal large language models (MLLMs) frequently hallucinate by over-committing to spurious visual cues. Prior remedies-Visual and Instruction Contrastive Decoding (VCD, ICD)-mitigate this issue, yet the mechanism remains opaque. We fir…
View article: FedNano: Toward Lightweight Federated Tuning for Pretrained Multimodal Large Language Models
FedNano: Toward Lightweight Federated Tuning for Pretrained Multimodal Large Language Models Open
Multimodal Large Language Models (MLLMs) excel in tasks like multimodal reasoning and cross-modal retrieval but face deployment challenges in real-world scenarios due to distributed multimodal data and strict privacy requirements. Federate…
View article: Agentic Neural Networks: Self-Evolving Multi-Agent Systems via Textual Backpropagation
Agentic Neural Networks: Self-Evolving Multi-Agent Systems via Textual Backpropagation Open
Leveraging multiple Large Language Models(LLMs) has proven effective for addressing complex, high-dimensional tasks, but current approaches often rely on static, manually engineered multi-agent configurations. To overcome these constraints…
View article: Improving LLM Reasoning through Interpretable Role-Playing Steering
Improving LLM Reasoning through Interpretable Role-Playing Steering Open
Role-playing has emerged as an effective technique for enhancing the reasoning capabilities of large language models (LLMs). However, existing methods primarily rely on prompt engineering, which often lacks stability and interpretability. …
View article: Language Mixing in Reasoning Language Models: Patterns, Impact, and Internal Causes
Language Mixing in Reasoning Language Models: Patterns, Impact, and Internal Causes Open
Reasoning language models (RLMs) excel at complex tasks by leveraging a chain-of-thought process to generate structured intermediate steps. However, language mixing, i.e., reasoning steps containing tokens from languages other than the pro…
View article: CoT-Kinetics: A Theoretical Modeling Assessing LRM Reasoning Process
CoT-Kinetics: A Theoretical Modeling Assessing LRM Reasoning Process Open
Recent Large Reasoning Models significantly improve the reasoning ability of Large Language Models by learning to reason, exhibiting the promising performance in solving complex tasks. LRMs solve tasks that require complex reasoning by exp…
View article: WebPilot: A Versatile and Autonomous Multi-Agent System for Web Task Execution with Strategic Exploration
WebPilot: A Versatile and Autonomous Multi-Agent System for Web Task Execution with Strategic Exploration Open
LLM-based autonomous agents often fail to execute complex web tasks that require dynamic interaction, largely due to the inherent uncertainty and complexity of these environments. Existing LLM-based web agents typically rely on rigid, expe…
View article: In-depth Analysis of Graph-based RAG in a Unified Framework
In-depth Analysis of Graph-based RAG in a Unified Framework Open
Graph-based Retrieval-Augmented Generation (RAG) has proven effective in integrating external knowledge into large language models (LLMs), improving their factual accuracy, adaptability, interpretability, and trustworthiness. A number of g…
View article: PRISM: Self-Pruning Intrinsic Selection Method for Training-Free Multimodal Data Selection
PRISM: Self-Pruning Intrinsic Selection Method for Training-Free Multimodal Data Selection Open
Visual instruction tuning adapts pre-trained Multimodal Large Language Models (MLLMs) to follow human instructions for real-world applications. However, the rapid growth of these datasets introduces significant redundancy, leading to incre…
View article: LLaVA Steering: Visual Instruction Tuning with 500x Fewer Parameters through Modality Linear Representation-Steering
LLaVA Steering: Visual Instruction Tuning with 500x Fewer Parameters through Modality Linear Representation-Steering Open
Multimodal Large Language Models (MLLMs) have significantly advanced visual tasks by integrating visual representations into large language models (LLMs). The textual modality, inherited from LLMs, equips MLLMs with abilities like instruct…
View article: PERFT: Parameter-Efficient Routed Fine-Tuning for Mixture-of-Expert Model
PERFT: Parameter-Efficient Routed Fine-Tuning for Mixture-of-Expert Model Open
The Mixture-of-Experts (MoE) paradigm has emerged as a powerful approach for scaling transformers with improved resource utilization. However, efficiently fine-tuning MoE models remains largely underexplored. Inspired by recent works on Pa…
View article: VideoINSTA: Zero-shot Long Video Understanding via Informative Spatial-Temporal Reasoning with LLMs
VideoINSTA: Zero-shot Long Video Understanding via Informative Spatial-Temporal Reasoning with LLMs Open
In the video-language domain, recent works in leveraging zero-shot Large Language Model-based reasoning for video understanding have become competitive challengers to previous end-to-end models. However, long video understanding presents u…
View article: WebPilot: A Versatile and Autonomous Multi-Agent System for Web Task Execution with Strategic Exploration
WebPilot: A Versatile and Autonomous Multi-Agent System for Web Task Execution with Strategic Exploration Open
LLM-based autonomous agents often fail to execute complex web tasks that require dynamic interaction due to the inherent uncertainty and complexity of these environments. Existing LLM-based web agents typically rely on rigid, expert-design…
View article: SA-DQAS: Self-attention Enhanced Differentiable Quantum Architecture Search
SA-DQAS: Self-attention Enhanced Differentiable Quantum Architecture Search Open
We introduce SA-DQAS, a novel framework that enhances Differentiable Quantum Architecture Search (DQAS) by integrating a self-attention mechanism, enabling more effective quantum circuit design for variational quantum algorithms. Unlike DQ…
View article: Quantum Architecture Search with Unsupervised Representation Learning
Quantum Architecture Search with Unsupervised Representation Learning Open
Unsupervised representation learning presents new opportunities for advancing Quantum Architecture Search (QAS) on Noisy Intermediate-Scale Quantum (NISQ) devices. QAS is designed to optimize quantum circuits for Variational Quantum Algori…
View article: Differentiable Quantum Architecture Search For Job Shop Scheduling Problem
Differentiable Quantum Architecture Search For Job Shop Scheduling Problem Open
The Job shop scheduling problem (JSSP) plays a pivotal role in industrial applications, such as signal processing (SP) and steel manufacturing, involving sequencing machines and jobs to maximize scheduling efficiency. Before, JSSP was solv…
View article: zrLLM: Zero-Shot Relational Learning on Temporal Knowledge Graphs with Large Language Models
zrLLM: Zero-Shot Relational Learning on Temporal Knowledge Graphs with Large Language Models Open
Modeling evolving knowledge over temporal knowledge graphs (TKGs) has become a heated topic. Various methods have been proposed to forecast links on TKGs. Most of them are embedding-based, where hidden representations are learned to repres…
View article: GraphextQA: A Benchmark for Evaluating Graph-Enhanced Large Language Models
GraphextQA: A Benchmark for Evaluating Graph-Enhanced Large Language Models Open
While multi-modal models have successfully integrated information from image, video, and audio modalities, integrating graph modality into large language models (LLMs) remains unexplored. This discrepancy largely stems from the inherent di…
View article: GenTKG: Generative Forecasting on Temporal Knowledge Graph with Large Language Models
GenTKG: Generative Forecasting on Temporal Knowledge Graph with Large Language Models Open
The rapid advancements in large language models (LLMs) have ignited interest in the temporal knowledge graph (tKG) domain, where conventional embedding-based and rule-based methods dominate. The question remains open of whether pre-trained…
View article: Differentiable Quantum Architecture Search for Quantum Reinforcement Learning
Differentiable Quantum Architecture Search for Quantum Reinforcement Learning Open
Differentiable quantum architecture search (DQAS) is a gradient-based framework to design quantum circuits automatically in the NISQ era. It was motivated by such as low fidelity of quantum hardware, low flexibility of circuit architecture…
View article: Benchmarking Robustness of Adaptation Methods on Pre-trained Vision-Language Models
Benchmarking Robustness of Adaptation Methods on Pre-trained Vision-Language Models Open
Various adaptation methods, such as LoRA, prompts, and adapters, have been proposed to enhance the performance of pre-trained vision-language models in specific domains. The robustness of these adaptation methods against distribution shift…
View article: QNEAT: Natural Evolution of Variational Quantum Circuit Architecture
QNEAT: Natural Evolution of Variational Quantum Circuit Architecture Open
Quantum Machine Learning (QML) is a recent and rapidly evolving field where the theoretical framework and logic of quantum mechanics are employed to solve machine learning tasks. Various techniques with different levels of quantum-classica…