Nanjing University

4 related articles

arXiv Sep 30, 2026
SaveRouter Cuts LLM Routing's Break-Even Deployment Volume by Up to About 9.5x Using Roughly a Third of the Supervision
The work proposes SaveRouter, a sparse-supervision routing framework that selectively acquires informative query-model feedback and shares capability information across related queries, using only about 33-41% of available training feedback across four routing benchmarks while maintaining competitive or better routing quality and reducing the break-even deployment volume by roughly 1.9-9.5x relative to the fastest conventional fully supervised router.
The work proposes SaveRouter, a sparse-supervision routing framework that selectively acquires informative query-model feedback and shares capability information across related queries, using only about 33-41% of available training feedback across four routing benchmarks while maintaining competitive or better routing quality and reducing the break-even deployment volume by roughly 1.9-9.5x relative to the fastest conventional fully supervised router.
Nanjing University
Sparse-Supervision Routing LLM Routing
ACM Transactions on Software Engineering and Methodology Sep 24, 2026
First systematic empirical study of diffusion LLMs for code generation: 7 diffusion models reach 79.1% on MBPP+ versus 73.3% for the best autoregressive baseline, yet open-source diffusion models still do not consistently match strong autoregressive models
This study empirically evaluates 7 representative diffusion LLMs (including the closed-source Mercury-Coder-Small) against 4 open-source autoregressive baselines on HumanEval/HumanEval+, MBPP/MBPP+, HumanEval-X, LiveCodeBench, RepoQA, RepoBench-C, and SWE-Bench Verified, finding promising but uneven code-generation ability: the closed-source diffusion model leads on most benchmarks (e.g., 79.1% on MBPP+ versus 73.
This study empirically evaluates 7 representative diffusion LLMs (including the closed-source Mercury-Coder-Small) against 4 open-source autoregressive baselines on HumanEval/HumanEval+, MBPP/MBPP+, HumanEval-X, LiveCodeBench, RepoQA, RepoBench-C, and SWE-Bench Verified, finding promising but uneven code-generation ability: the closed-source diffusion model leads on most benchmarks (e.g., 79.1% on MBPP+ versus 73.
Nanjing University Peking University Tsinghua University Jia Li Ge Li Yitong Zhang Liyi Cai Chengze Li
Topic Modeling Natural Language Processing Techniques
arXiv Sep 25, 2026
Nanjing University team releases DSEBench, the first test collection supporting both keyword-plus-example dataset search and field-level explanations
The work formalizes Dataset Search with Examples (DSE) and its explainable extension (ExDSE), builds DSEBench as the first test collection with both dataset-level and field-level human annotations (141 test cases, 7,415 triples, plus 5,699 training cases and 122,585 triples), uses a large language model to generate large-scale training annotations, and adapts and evaluates a range of retrieval, reranking, and explanation methods to establish baselines.
The work formalizes Dataset Search with Examples (DSE) and its explainable extension (ExDSE), builds DSEBench as the first test collection with both dataset-level and field-level human annotations (141 test cases, 7,415 triples, plus 5,699 training cases and 122,585 triples), uses a large language model to generate large-scale training annotations, and adapts and evaluates a range of retrieval, reranking, and explanation methods to establish baselines.
Nanjing University Nanjing University of Chinese Medicine Jing He Gong Cheng Qiaosheng Chen Qing Shi
Topic Modeling Semantic Web and Ontologies Information Retrieval and Search Behavior
arXiv Sep 16, 2026
Re-evaluating the Advancements of Heterophilic Graph Learning: A Three-Way Dataset Taxonomy and Quantitative Evaluation of Homophily Metrics
This work fine-tunes baseline models on 27 widely used benchmark datasets, categorizes heterophilic datasets into malignant, benign, and ambiguous groups based on whether graph-aware models underperform their coupled graph-agnostic counterparts, re-evaluates 11 SOTA heterophily-specific models, and conducts the first quantitative evaluation of 11 homophily metrics on synthetic graphs from three generation methods, finding that most SOTA models do not significantly outperform the best baselines and that classic metrics remain competitive.
This work fine-tunes baseline models on 27 widely used benchmark datasets, categorizes heterophilic datasets into malignant, benign, and ambiguous groups based on whether graph-aware models underperform their coupled graph-agnostic counterparts, re-evaluates 11 SOTA heterophily-specific models, and conducts the first quantitative evaluation of 11 homophily metrics on synthetic graphs from three generation methods, finding that most SOTA models do not significantly outperform the best baselines and that classic metrics remain competitive.
Université de Montréal McGill University Nanjing University Nvidia (United States)Mila - Quebec Artificial Intelligence Institute Chenqing Hua Qincheng Lu Jiaqi Zhu Xinyu Wang Sitao Luan Xiao-Wen Chang
Advanced Graph Neural Networks Domain Adaptation and Few-Shot Learning Machine Learning in Healthcare

You’re all caught up

SaveRouter Cuts LLM Routing's Break-Even Deployment Volume by Up to About 9.5x Using Roughly a Third of the Supervision

First systematic empirical study of diffusion LLMs for code generation: 7 diffusion models reach 79.1% on MBPP+ versus 73.3% for the best autoregressive baseline, yet open-source diffusion models still do not consistently match strong autoregressive models

Nanjing University team releases DSEBench, the first test collection supporting both keyword-plus-example dataset search and field-level explanations

Re-evaluating the Advancements of Heterophilic Graph Learning: A Three-Way Dataset Taxonomy and Quantitative Evaluation of Homophily Metrics