Built with and by Teycir Ben Soltane•
How to Use•FAQ•GitHub•arXiv.org•
Share:
ArXivCSExplorer
☆☆Bookmarks🏆RSSHow to UseFAQ
Home/Authors/Wang Yang

Wang Yang

16 indexed papers

Recent (6 mo)
15
With code
0
Influential cites
0
Benchmarked
0

Publications per year

16
26

Top categories

AI×10NLP×5Vision×3Crypto×2Sound×2Info Retrieval×1ML×1Multiagent×1

Frequent co-authors

Yang Yang4×
Yang Wang4×
Chuang Ma3×
Qianying Liu3×
Yang Liu2×
Xiang Li2×

Research Timeline

2026
MOSAIC: Multi-Objective Slice-Aware Iterative Curation for Alignment

MOSAIC is a multi-objective framework that efficiently allocates a fixed supervised fine-tuning budget by turning failure profiles into actionable data mixtures, significantly improving model alignment across multiple safety and capability metrics.

FedMPT: Federated Multi-label Prompt Tuning of Vision-Language Models

FedMPT introduces a novel federated learning framework for Multi-Label Recognition (MLR) using Vision-Language Models (VLMs) by leveraging generalizable conditions to mitigate label overfitting and improve robustness.

FedMPT: Federated Multi-label Prompt Tuning of Vision-Language Models

FedMPT introduces a novel federated learning framework for Multi-Label Recognition (MLR) using Vision-Language Models (VLMs) by leveraging generalizable conditions to mitigate label overfitting and improve robustness.

PTCG-Bench: Can LLM Agents Master Pokémon Trading Card Game?

The paper introduces PTCG-Bench, a new benchmark using the Pokémon TCG to evaluate LLM agents' strategic decision-making and ability to self-evolve, finding that sustained self-evolution remains challenging.

PTCG-Bench: Can LLM Agents Master Pokémon Trading Card Game?

The paper introduces PTCG-Bench, a new benchmark using the Pokémon TCG to evaluate LLM agents' strategic decision-making and ability to self-evolve, finding that sustained self-evolution remains challenging.

eMoT: evolving Memory-of-Thought via Symbolic Anchoring and Memory Corrosion

The eMoT framework enhances multi-step reasoning in LLMs by treating reasoning as an evolving memory, stabilizing performance through symbolic computation and structured refinement.

MOSS-Audio Technical Report

MOSS-Audio is a unified audio-language model designed for comprehensive understanding of speech, environmental sounds, and music, achieving strong performance across various audio-grounded tasks.

Mechanistic Diagnostics of Spatial Lexical Bias in Multimodal Large Language Model Spatial Reasoning

The paper identifies a failure mode called spatial lexical bias in MLLMs, where adding a spatial word to options biases the model's choice, and demonstrates that this failure originates primarily from the language processing side rather than poor visual attention.

eMoT: evolving Memory-of-Thought via Symbolic Anchoring and Memory Corrosion

The eMoT framework enhances multi-step reasoning in LLMs by treating reasoning as an evolving memory, stabilizing performance through symbolic computation and structured refinement.

MOSS-Audio Technical Report

MOSS-Audio is a unified audio-language model designed for comprehensive understanding of speech, environmental sounds, and music, achieving strong performance across various audio-grounded tasks.

Mechanistic Diagnostics of Spatial Lexical Bias in Multimodal Large Language Model Spatial Reasoning

The paper identifies a failure mode called spatial lexical bias in MLLMs, where adding a spatial word to options biases the model's choice, and demonstrates that this failure originates primarily from the language processing side rather than poor visual attention.

Selection Integrity for LLM Graph Memory: An Accumulability Criterion for Information-Flow-Blind Retrieval

One clear sentence describing what this paper does

ARP: Enhancing Quantized Skill Abstractions via Visual Alignment and Iterative Refinement for Robotic Manipulation

This paper proposes a new framework, Aligned Refinement Policy (ARP), to learn visuomotor policies for long-horizon manipulation tasks.

O-VAD: Industrial Video Anomaly Detection through Object-Centric Tracking and Reasoning

This paper introduces an agentic framework for industrial video anomaly detection, capable of tracking spatial-temporal dynamics and underlying transformations of objects to identify abnormalities.

How to Loop MoE: Flatten the Experts, Untie the Attention

The paper presents Foil, a method for looping sparse mixture-of-experts (MoE) models to improve expert usage and performance.

SLIDE: Search-to-List Item Distillation for Single-Rollout Generative Retrieval

The paper proposes a search-to-list item distillation framework called SLIDE to improve generative retrieval by decoupling offline search from online list generation.

Highlighted terms show continued research focus across papers

Papers

cs.IREmpiricalRecentOct 26, 2026

SLIDE: Search-to-List Item Distillation for Single-Rollout Generative Retrieval

Tianyu Zhan, Gui Ling, Ran Tao, Kunhai Lin +8 more

The paper proposes a search-to-list item distillation framework called SLIDE to improve generative retrieval by decoupling offline search from online list generation.

View →
cs.LGcs.AIcs.CLEmpirical
Recent
Sep 28, 2026

How to Loop MoE: Flatten the Experts, Untie the Attention

Shouren Wang, Chuang Ma, Mohsen Hariri, Debargha Ganguly +5 more

The paper presents Foil, a method for looping sparse mixture-of-experts (MoE) models to improve expert usage and performance.

View →
cs.CVcs.AIcs.CLEmpiricalRecentJul 20, 2026

O-VAD: Industrial Video Anomaly Detection through Object-Centric Tracking and Reasoning

Mei Yuan, Qi Long, Qifeng Wu, Zhenyang Li +4 more

This paper introduces an agentic framework for industrial video anomaly detection, capable of tracking spatial-temporal dynamics and underlying transformations of objects to identify abnormalities.

View →
cs.ROEmpiricalRecentJun 21, 2026

ARP: Enhancing Quantized Skill Abstractions via Visual Alignment and Iterative Refinement for Robotic Manipulation

Yuntian Wang, Zesheng Jia, Yuhui Duan, Qibing Wang +4 more

This paper proposes a new framework, Aligned Refinement Policy (ARP), to learn visuomotor policies for long-horizon manipulation tasks.

View →
cs.CREmpiricalRecentJun 10, 2026

Selection Integrity for LLM Graph Memory: An Accumulability Criterion for Information-Flow-Blind Retrieval

Zeming Fei, Hongming Fei, Xiaoyang Wang, Yang yang +3 more

One clear sentence describing what this paper does

View →
cs.AIRecentJun 1, 2026

eMoT: evolving Memory-of-Thought via Symbolic Anchoring and Memory Corrosion

Xiang Li, Jiwei Wei, Ke Liu, Yitong Qin +4 more

The eMoT framework enhances multi-step reasoning in LLMs by treating reasoning as an evolving memory, stabilizing performance through symbolic computation and structured refinement.

View →
cs.SDcs.AIRecentJun 1, 2026

MOSS-Audio Technical Report

Chen Yang, Chufan Yu, Hanfu Chen, Jie Zhu +21 more

MOSS-Audio is a unified audio-language model designed for comprehensive understanding of speech, environmental sounds, and music, achieving strong performance across various audio-grounded tasks.

View →
cs.CLcs.CVRecentJun 1, 2026

Mechanistic Diagnostics of Spatial Lexical Bias in Multimodal Large Language Model Spatial Reasoning

Chuang Ma, Qianying Liu, Tomoyuki Obuchi, Fei Cheng +5 more

The paper identifies a failure mode called spatial lexical bias in MLLMs, where adding a spatial word to options biases the model's choice, and demonstrates that this failure originates primarily from…

View →
cs.AIRecentJun 1, 2026

eMoT: evolving Memory-of-Thought via Symbolic Anchoring and Memory Corrosion

Xiang Li, Jiwei Wei, Ke Liu, Yitong Qin +4 more

The eMoT framework enhances multi-step reasoning in LLMs by treating reasoning as an evolving memory, stabilizing performance through symbolic computation and structured refinement.

View →
cs.SDcs.AIRecentJun 1, 2026

MOSS-Audio Technical Report

Chen Yang, Chufan Yu, Hanfu Chen, Jie Zhu +21 more

MOSS-Audio is a unified audio-language model designed for comprehensive understanding of speech, environmental sounds, and music, achieving strong performance across various audio-grounded tasks.

View →
cs.CLcs.CVRecentJun 1, 2026

Mechanistic Diagnostics of Spatial Lexical Bias in Multimodal Large Language Model Spatial Reasoning

Chuang Ma, Qianying Liu, Tomoyuki Obuchi, Fei Cheng +5 more

The paper identifies a failure mode called spatial lexical bias in MLLMs, where adding a spatial word to options biases the model's choice, and demonstrates that this failure originates primarily from…

View →
cs.AIRecentMay 28, 2026

PTCG-Bench: Can LLM Agents Master Pokémon Trading Card Game?

Dongdong Hua, Yifei Sun, Renhong Huang, Feng Gao +2 more

The paper introduces PTCG-Bench, a new benchmark using the Pokémon TCG to evaluate LLM agents' strategic decision-making and ability to self-evolve, finding that sustained self-evolution remains chall…

View →
cs.AIRecentMay 28, 2026

PTCG-Bench: Can LLM Agents Master Pokémon Trading Card Game?

Dongdong Hua, Yifei Sun, Renhong Huang, Feng Gao +2 more

The paper introduces PTCG-Bench, a new benchmark using the Pokémon TCG to evaluate LLM agents' strategic decision-making and ability to self-evolve, finding that sustained self-evolution remains chall…

View →
cs.AIRecentMay 27, 2026

FedMPT: Federated Multi-label Prompt Tuning of Vision-Language Models

Xucong Wang, Pengkun Wang, Zhe Zhao, Liheng Yu +2 more

FedMPT introduces a novel federated learning framework for Multi-Label Recognition (MLR) using Vision-Language Models (VLMs) by leveraging generalizable conditions to mitigate label overfitting and im…

View →
cs.AIRecentMay 27, 2026

FedMPT: Federated Multi-label Prompt Tuning of Vision-Language Models

Xucong Wang, Pengkun Wang, Zhe Zhao, Liheng Yu +2 more

FedMPT introduces a novel federated learning framework for Multi-Label Recognition (MLR) using Vision-Language Models (VLMs) by leveraging generalizable conditions to mitigate label overfitting and im…

View →
cs.CRcs.CLMar 19, 2026

MOSAIC: Multi-Objective Slice-Aware Iterative Curation for Alignment

Yipu Dou, Wang Yang

MOSAIC is a multi-objective framework that efficiently allocates a fixed supervised fine-tuning budget by turning failure profiles into actionable data mixtures, significantly improving model alignmen…

View →