Bolin Ding

Senior Director@Alibaba Group·美国·OpenReview

研究方向

recommendation system · automl · BERT · approximate query processing · sampling · data privacy · differential privacy · optimization · approximation algorithm

Bolin Ding

Beyond Magnitude: Leveraging Direction of RLVR Updates for LLM Reasoning

On-Policy RL Meets Off-Policy Experts: Harmonizing Supervised Fine-Tuning and Reinforcement Learning via Dynamic Weighting

HumanVBench: Probing Human-Centric Video Understanding in MLLMs with Automatically Synthesized Benchmarks

Sparse but Critical: A Token-Level Analysis of Distributional Shifts in RLVR Fine-Tuning of LLMs

Group-Relative REINFORCE Is Secretly an Off-Policy Algorithm: Demystifying Some Myths About GRPO and Its Friends

Grounded in Reality: Learning and Deploying Proactive LLM from Offline Logs

BOTS: A Unified Framework for Bayesian Online Task Selection in LLM Reinforcement Finetuning

CUES: Bottom-Up Exploration and Top-Down Guidance for Agentic Data Synthesis

Auto-Rubric: Learning to Extract Generalizable Criteria for Reward Modeling

MixLLM: Selecting Large Language Models with High-Quality Results and Minimum Inference Cost for Multi-Stage Complex Tasks

Taming the Judge: Deconflicting AI Feedback for Stable Reinforcement Learning

Omni-SafetyBench: A Benchmark for Safety Evaluation of Audio-Visual Large Language Models

What is Wrong with Perplexity for Long-context Language Modeling?

Provable Scaling Laws for the Test-Time Compute of Large Language Models

RePO: Understanding Preference Learning Through ReLU-Based Optimization

Towards Robust Alignment of Language Models: Distributionally Robustifying Direct Preference Optimization

Data-Juicer Sandbox: A Feedback-Driven Suite for Multimodal Data-Model Co-development

Data-Juicer Sandbox: A Comprehensive Suite for Multimodal Data-Model Co-development

Agent-Oriented Planning in Multi-Agent Systems

Larger or Smaller Reward Margins to Select Preferences for LLM Alignment?

Learning Bayesian Nash Equilibrium in Auction Games via Approximate Best Response

$\alpha$-DPO: Adaptive Reward Margin is What Direct Preference Optimization Needs

Make LLMs better zero-shot reasoners: structure-oriented autonomous reasoning

Large-Scale Dynamic Graph Generation via LLM-based Agent Simulation

Leveraging Diffusion Transformers for Stock Factor Augmentation in Financial Markets

BiMix: Bivariate Data Mixing Law for Language Model Pretraining

Exploring Selective Layer Fine-Tuning in Federated Learning

AlphaDPO: Adaptive Reward Margin for Direct Preference Optimization

Achieving Exact Federated Unlearning with Improved Post-Unlearning Performance

On the Design and Analysis of LLM-Based Algorithms

Very Large-Scale Multi-Agent Simulation with LLM-Powered Agents