14 papers
ANDES: Agent Native Data Evolving Synthesis Tool for Autonomous Instruction Alignment
Zhengyang Zhao, Shengjie Ye, Lu Ma +3
AI agents are increasingly being tasked with automating AI research itself, particularly the critical post-training phase that transforms base LLMs into aligned assistants. However…
Training with Harnesses: On-Policy Harness Self-Distillation for Complex Reasoning
Zhengyang Zhao, Lu Ma, Wentao Zhang
Inference-time harnesses substantially improve large language models on complex reasoning tasks. However, the intrinsic capabilities of the underlying model remain unchanged by the…
Fractal Characterization of Low-Correlation Signals in AI-Generated Image Detection
Wenwei Xie, Jie Yin, Lu Ma +2
AI-generated imagery has reached near-photorealistic fidelity, yet this technology poses significant threats to information security and societal trust. Existing deepfake detection…
DataFlex: A Unified Framework for Data-Centric Dynamic Training of Large Language Models
Hao Liang, Zhengyang Zhao, Meiyi Qiang +22
Data-centric training has emerged as a promising direction for improving large language models (LLMs) by optimizing not only model parameters but also the selection, composition, a…
Thinking by Subtraction: Confidence-Driven Contrastive Decoding for LLM Reasoning
Lexiang Tang, Weihao Gao, Bingchen Zhao +4
Recent work on test-time scaling for large language model (LLM) reasoning typically assumes that allocating more inference-time computation uniformly improves correctness. However,…
GIFT: Reconciling Post-Training Objectives via Variational Finite-Temperature Gibbs Initialization
Zhengyang Zhao, Lu Ma, Yizhen Jiang +7
The prevailing post-training paradigm for Large Reasoning Models (LRMs)---Supervised Fine-Tuning (SFT) followed by Reinforcement Learning (RL)-suffers from an intrinsic optimizatio…
Leash: Adaptive Length Penalty and Reward Shaping for Efficient Large Reasoning Model
Yanhao Li, Lu Ma, Jiaran Zhang +3
Existing approaches typically rely on fixed length penalties, but such penalties are hard to tune and fail to adapt to the evolving reasoning abilities of LLMs, leading to suboptim…
DataFlow: An LLM-Driven Framework for Unified Data Preparation and Workflow Automation in the Era of Data-Centric AI
Hao Liang, Xiaochen Ma, Zhou Liu +32
The rapidly growing demand for high-quality data in Large Language Models (LLMs) has intensified the need for scalable, reliable, and semantically rich data preparation pipelines.…
DARO: Difficulty-Aware Reweighting Policy Optimization
Jingyu Zhou, Lu Ma, Hao Liang +3
Recent advances in large language models (LLMs) have shown that reasoning ability can be significantly enhanced through Reinforcement Learning with Verifiable Rewards (RLVR). Group…
Ming-Flash-Omni: A Sparse, Unified Architecture for Multimodal Perception and Generation
Inclusion AI, :, Bowen Ma +73
We propose Ming-Flash-Omni, an upgraded version of Ming-Omni, built upon a sparser Mixture-of-Experts (MoE) variant of Ling-Flash-2.0 with 100 billion total parameters, of which on…
Not All Tokens and Heads Are Equally Important: Dual-Level Attention Intervention for Hallucination Mitigation
Lexiang Tang, Xianwei Zhuang, Bang Yang +5
Large vision-language models (LVLMs) have demonstrated impressive capabilities across diverse multimodal tasks, yet they remain highly susceptible to visual hallucinations (VH), of…
Learning What Reinforcement Learning Can't: Interleaved Online Fine-Tuning for Hardest Questions
Lu Ma, Hao Liang, Meiyi Qiang +9
Recent advances in large language model (LLM) reasoning have shown that sophisticated behaviors such as planning and self-reflection can emerge through reinforcement learning (RL).…
Evaluating and Predicting Distorted Human Body Parts for Generated Images
Lu Ma, Kaibo Cao, Hao Liang +6
Recent advancements in text-to-image (T2I) models enable high-quality image synthesis, yet generating anatomically accurate human figures remains challenging. AI-generated images f…
Acceleration Algorithms in GNNs: A Survey
Lu Ma, Zeang Sheng, Xunkai Li +5
Graph Neural Networks (GNNs) have demonstrated effectiveness in various graph-based tasks. However, their inefficiency in training and inference presents challenges for scaling up…