paper-with-me

Papers

SPARK: Self-Play with Asymmetric Reward from Knowledge Graphs

2026-05-07 · Hyobin Park, Taeseop Kim, Dong-Geol Choi arxiv

Self-play reinforcement learning has shown strong performance in domains with formally verifiable structure, such as mathematics and coding, where both problem generation and reward computation can be grounded in explicit rules. Extending this paradigm to scientific literature is more challenging: the relationships among multi-modal elements within and across documents are rarely made explicit in text, which makes automatic generation of relational reasoning questions difficult and weakens the reliability of reward signals. We propose SPARK (Self-Play with Asymmetric Reward from Knowledge Graphs), a framework that automatically constructs a unified knowledge graph (KG) from multi-document scientific literature and uses it as the structural basis for self-play. KG paths over multimodal nodes serve as a source for generating relational reasoning questions, and structured facts stored in the KG provide a basis for verifiable reward computation. A single small vision-language model (sVLM) alternates between Proposer and Solver roles under information asymmetry against a fixed KG, a design that we believe can be naturally extended toward online adaptation in future work. We evaluate SPARK on public benchmarks and a self-constructed cross-document multi-hop QA dataset. Results show that SPARK consistently outperforms flat-corpus-based self-play baselines, and the performance gap widens as hop count increases, suggesting that KG-structure grounding contributes to relational multi-hop reasoning beyond what unstructured corpus grounding can provide.

📄 PDF Abstract BibTeX arXiv:2605.05546

Code (0)

등록된 구현이 없습니다.

Tasks

Reinforcement LearningRelational ReasoningKnowledge Graphs

Similar Papers 제목 키워드 기반

Survive or Collapse: The Asymmetric Roles of Data Gating and Reward Grounding in Self-Play RL

2026-05-21 · Sophia Xiao Pu, Zhaotian Weng, Chengzhi Liu, Jayanth Srinivasa 외 arxiv

Self-play reinforcement learning trains language models on their own generated tasks, co-evolving a proposer and solver without human labels. Recent systems report strong reasoning gains, but collapse and instability are…

Reinforcement Learning

Asymmetric self-play for automatic goal discovery in robotic manipulation

2021-01-13 · OpenAI OpenAI, Matthias Plappert, Raul Sampedro, Tao Xu 외

We train a single, goal-conditioned policy that can solve many robotic manipulation tasks, including tasks with previously unseen goals and objects. We rely on asymmetric self-play for goal discovery, where two agents, A…

SPARK: Synergistic Policy And Reward Co-Evolving Framework

2025-09-26 · Ziyu Liu, Yuhang Zang, Shengyuan Ding, Yuhang Cao 외 arxiv

Recent Large Language Models (LLMs) and Large Vision-Language Models (LVLMs) increasingly use Reinforcement Learning (RL) for post-pretraining, such as RL with Verifiable Rewards (RLVR) for objective tasks and RL from Hu…

Reinforcement Learning

Self-Questioning Language Models

2025-08-05 · Lili Chen, Mihir Prabhudesai, Katerina Fragkiadaki, Hao Liu 외 arxiv

Can large language models improve without external data -- by generating their own questions and answers? We hypothesize that a pre-trained language model can improve its reasoning skills given only a single prompt speci…

Reinforcement Learning

DARC: Decoupled Asymmetric Reasoning Curriculum for LLM Evolution

2026-01-20 · Shengda Fan, Xuyan Ye, Yankai Lin arxiv

Self-play with large language models has emerged as a promising paradigm for achieving self-improving artificial intelligence. However, existing self-play frameworks often suffer from optimization instability, due to (i)…