Natural Language Processing

URSA-7B: A new multimodal model significantly improves chain-of-thought reasoning in mathematics!

Small language models can master complex math reasoning using self-evolved deep thinking via Monte Carlo Tree Search, surpassing larger models in performance.

LLMs revolutionize scientific research! This survey reveals their transformative potential across hypothesis discovery, experiment planning, writing, and peer review, guiding future research.

EpiCoder revolutionizes code generation by using feature trees to create diverse and complex training data, resulting in state-of-the-art performance on various benchmarks.

First-ever resources (NER dataset, dependency treebank, and corpus) and models for historical Turkish NLP are introduced, significantly advancing research capabilities in this underexplored field.

PPTAgent, a novel two-stage framework, significantly improves automatic presentation generation by leveraging an edit-based workflow and a new evaluation metric, outperforming existing end-to-end meth…

Boosting private LLMs’ efficiency and security, this research introduces an entropy-guided attention mechanism and PI-friendly layer normalization to mitigate the overheads of nonlinear operations.

Samba-ASR, a novel speech recognition model using Mamba architecture, surpasses existing transformer models in accuracy and efficiency, setting a new benchmark for future ASR research.

GeAR, a new retrieval model, boosts accuracy by combining document retrieval with fine-grained information generation, leading to better understanding and improved localization.

BoostStep enhances large language models’ mathematical abilities by refining single-step reasoning through a novel step-level in-context learning strategy, achieving significant improvements on variou…

ToolHop: New benchmark dataset rigorously evaluates LLMs’ multi-hop tool use, revealing significant challenges and variations across different LLM families.

Unlocking LLM potential: This paper surveys test-time computing, showing how it boosts reasoning abilities by shifting from reactive System-1 to deliberate System-2 thinking, paving the way for more p…

New scaling laws for efficient floating-point quantization training in LLMs are presented, showing optimal bit allocation and critical data size.

REINFORCE++, a novel RLHF algorithm, achieves superior training stability and computational efficiency compared to existing methods like PPO and GRPO, while maintaining comparable performance.

Personalized Graph-based Retrieval-Augmented Generation (PGraphRAG) significantly improves personalized text generation by leveraging user-centric knowledge graphs, especially in cold-start scenarios …

METAGENE-1, a 7-billion parameter language model, achieves state-of-the-art results in pathogen detection and genomic embedding by leveraging a massive wastewater metagenomic dataset.

AUTO-RT automates LLM vulnerability discovery by using reinforcement learning to optimize complex attack strategies, achieving faster detection and higher success rates than existing methods.

Boosting code generation accuracy with more unit tests! This research shows that increasing the number of unit tests used to evaluate code generated by LLMs significantly improves accuracy, especially…

CODEELO benchmark uses CodeForces to fairly evaluate LLMs’ coding abilities, providing human-comparable Elo ratings and addressing limitations of existing benchmarks.

BoxingGym: A new benchmark rigorously evaluates AI agents’ ability to design experiments and discover scientific models, revealing current LLMs’ limitations and highlighting fertile research avenues.