Learning Like Humans: Advancing LLM Reasoning Capabilities via Adaptive Difficulty Curriculum Learning and Expert-Guided Self-Reformulation
Despite impressive progress in areas like mathematical reasoning, large language models still face significant challenges in consistently solving complex problems. Drawing inspiration from key human learning strategies, we propose two novel strategies to enhance the capability of large language models to solve these complex problems. First, Adaptive Difficulty Curriculum Learning (ADCL) is a novel curriculum learning strategy that tackles the Difficulty Shift phenomenon (i.e., a model's perception of problem difficulty dynamically changes during training) by periodically re-estimating difficulty within upcoming data batches to maintain alignment with the model's evolving capabilities. Second, Expert-Guided Self-Reformulation (EGSR) is a novel reinforcement learning strategy that bridges the gap between imitation learning and pure exploration by guiding models to reformulate expert solutions within their own conceptual framework, rather than relying on direct imitation, fostering deeper understanding and knowledge assimilation. Extensive experiments on challenging mathematical reasoning benchmarks, using Qwen2.5-7B as the base model, demonstrate that these human-inspired strategies synergistically and significantly enhance performance. Notably, their combined application improves performance over the standard Zero-RL baseline by 10% on the AIME24 benchmark and 16.6% on AIME25.
Code (0)
등록된 구현이 없습니다.
Tasks
Imitation LearningMathematical ReasoningMethods 이 논문이 사용한 방법론
Similar Papers 제목 키워드 기반
11Plus-Bench: Demystifying Multimodal LLM Spatial Reasoning with Cognitive-Inspired Analysis
For human cognitive process, spatial reasoning and perception are closely entangled, yet the nature of this interplay remains underexplored in the evaluation of multimodal large language models (MLLMs). While recent MLLM…
Spatial ReasoningTime-Scaling Is What Agents Need Now
Early artificial intelligence paradigms exhibited separated cognitive functions: Neural Networks focused on "perception-representation," Reinforcement Learning on "decision-making-behavior," and Symbolic AI on "knowledge…
Reinforcement LearningReasoningV: Efficient Verilog Code Generation with Adaptive Hybrid Reasoning Model
Large Language Models (LLMs) have advanced Verilog code generation significantly, yet face challenges in data quality, reasoning capabilities, and computational efficiency. This paper presents ReasoningV, a novel model e…
Code GenerationComputational Efficiencyparameter-efficient fine-tuningWorld Models in Artificial Intelligence: Sensing, Learning, and Reasoning Like a Child
World Models help Artificial Intelligence (AI) predict outcomes, reason about its environment, and guide decision-making. While widely used in reinforcement learning, they lack the structured, adaptive representations th…
Causal InferenceContinual LearningDecision MakingMM-MATH: Advancing Multimodal Math Evaluation with Process Evaluation and Fine-grained Classification
To advance the evaluation of multimodal math reasoning in large multimodal models (LMMs), this paper introduces a novel benchmark, MM-MATH. MM-MATH consists of 5,929 open-ended middle school math problems with visual con…
Image ComprehensionMathMultimodal Reasoning