AdaTok learns content-dependent token budgets for discrete 1D image tokenization via prioritized representation learning and a GRPO allocation policy, achieving rFID 1.50 at ~118 tokens average versus fixed 256-token baselines.
Learning to optimize multi-objective alignment through dynamic reward weighting
7 Pith papers cite this work. Polarity classification is still indexing.
7
Pith papers citing it
citation-role summary
background 1
citation-polarity summary
years
2026 7verdicts
UNVERDICTED 7roles
background 1polarities
background 1representative citing papers
LEAD uses online adaptive mechanisms including Potential-Scaled Instability and symmetric efficiency rewards based on correct rollouts to achieve higher accuracy-efficiency scores with substantially shorter reasoning outputs than base models on math benchmarks.
ARL-RR alternates optimization over rubric meta-classes with dynamic selection to avoid fixed scalarization, outperforming baselines on HealthBench.
POW3R adapts rubric criterion weights via rollout contrast in RLVR to improve mean reward, strict completion rates, and training speed over static rubric aggregation on multimodal and text tasks.
MP2D is a framework that guides discrete diffusion denoising with constrained MCTS and Pareto rewards to optimize protein sequences for four to five simultaneous objectives, outperforming baselines on antimicrobial peptide and binder design tasks.
MAMO uses multi-agent RL to automatically select reward weights for constrained optimization problems in non-stationary environments.
Taiji presents a LLM-as-Enhancer system with reverse-engineered CoT data generation and Pareto Optimal Policy Optimization (POPO) to trade off semantic and ID rewards, deployed at Kuaishou serving 400M daily users.
citing papers explorer
-
AdaTok: Self-Budgeting Image Tokenization with Quality-Preserving Dynamic Tokens
AdaTok learns content-dependent token budgets for discrete 1D image tokenization via prioritized representation learning and a GRPO allocation policy, achieving rFID 1.50 at ~118 tokens average versus fixed 256-token baselines.
-
LEAD: Length-Efficient Adaptive and Dynamic Reasoning for Large Language Models
LEAD uses online adaptive mechanisms including Potential-Scaled Instability and symmetric efficiency rewards based on correct rollouts to achieve higher accuracy-efficiency scores with substantially shorter reasoning outputs than base models on math benchmarks.
-
Alternating Reinforcement Learning with Contextual Rubric Rewards: Beyond the Scalarization Strategy
ARL-RR alternates optimization over rubric meta-classes with dynamic selection to avoid fixed scalarization, outperforming baselines on HealthBench.
-
Not Every Rubric Teaches Equally: Policy-Aware Rubric Rewards for RLVR
POW3R adapts rubric criterion weights via rollout contrast in RLVR to improve mean reward, strict completion rates, and training speed over static rubric aggregation on multimodal and text tasks.
-
MP2D: Constrained Monte Carlo Tree-Guided Diffusion for Multi-Objective Protein Sequence Design
MP2D is a framework that guides discrete diffusion denoising with constrained MCTS and Pareto rewards to optimize protein sequences for four to five simultaneous objectives, outperforming baselines on antimicrobial peptide and binder design tasks.
-
A Multi-Agent system for Multi-Objective constrained optimization
MAMO uses multi-agent RL to automatically select reward weights for constrained optimization problems in non-stationary environments.
-
Taiji: Pareto Optimal Policy Optimization with Semantics-IDs Trade-off for Industrial LLM-Enhanced Recommendation
Taiji presents a LLM-as-Enhancer system with reverse-engineered CoT data generation and Pareto Optimal Policy Optimization (POPO) to trade off semantic and ID rewards, deployed at Kuaishou serving 400M daily users.