Author name cluster

Yao Cheng

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

5 papers

2 author rows

AAAI Conference 2026 Conference Paper

Human Cognition Inspired RAG with Knowledge Graph for Complex Problem Solving

Yao Cheng
Yibo Zhao
Jiapeng Zhu
Yao Liu
Xing Sun
Xiang Li

Large Language Models (LLMs) have demonstrated significant potential across various domains. However, they often struggle with integrating external knowledge and performing complex reasoning, leading to hallucinations and unreliable outputs. Retrieval Augmented Generation (RAG) has emerged as a promising paradigm to mitigate these issues by incorporating external knowledge. Yet, conventional RAG approaches, especially those based on vector similarity, fail to effectively capture relational dependencies and support multi-step reasoning. In this work, we propose CogGRAG, a human cognition-inspired, graph-based RAG framework designed for Knowledge Graph Question Answering (KGQA). CogGRAG models the reasoning process as a tree-structured mind map that decomposes the original problem into interrelated subproblems and explicitly encodes their semantic relationships. This structure not only provides a global view to guide subsequent retrieval and reasoning but also enables self-consistent verification across reasoning paths. The framework operates in three stages: (1) top-down problem decomposition via mind map construction, (2) structured retrieval of both local and global knowledge from external Knowledge Graphs (KGs), and (3) bottom-up reasoning with dual-process self-verification. Unlike previous tree-based decomposition methods such as MindMap or Graph-CoT, CogGRAG unifies problem decomposition, knowledge retrieval, and reasoning under a single graph-structured cognitive framework, allowing early integration of relational knowledge and adaptive verification. Extensive experiments demonstrate that CogGRAG achieves superior accuracy and reliability compared to existing methods.

PDF Details DOI

EAAI Journal 2026 Journal Article

Nonlinear multi-field coupling analysis and prediction of thermo-mechanical response of oil shale under microwave heating using physics-informed neural network

Yao Cheng
Yinlong Zhu
Yulin Ma

Strongly nonlinear multi-field coupling problems in the microwave heating of oil shale require high-precision and efficient solution methods. Physics-informed neural network (PINN) has emerged as a powerful tool for solving partial differential equations (PDEs); however, existing PINN approaches often have limited applicability under specific conditions, and their training convergence remains suboptimal when applied to strongly nonlinear, temperature-dependent material problems. To address this, this study proposes a PINN framework with a strong physical constraint weighting strategy, in which the physical loss weight is constrained within 0. 5-0. 6 and dynamically balanced during training. Coupled with experimentally obtained temperature-dependent material parameters, this enables efficient surrogate solutions of the coupled PDEs. Three test cases (600 W, 800 W microwave heating, and cross-condition high-temperature steam heating of oil shale) were used for training and prediction. The results show that the PINN framework can output temperature and displacement fields at arbitrary points in milliseconds, with total loss stabilized at the 10−4-10−3 level, high prediction accuracy, and strong agreement with COMSOL simulations. These findings demonstrate that the proposed framework achieves high-precision, rapid prediction and cross-condition generalization for strongly nonlinear thermo-mechanical coupling problems in temperature-dependent materials.

Details DOI

YNIMG Journal 2025 Journal Article

Continuous theta-burst stimulation demonstrates language-network-specific causal effects on syntactic processing

Chenyang Gao
Junjie Wu
Yao Cheng
Yuming Ke
Xingfang Qu
Mingchuan Yang
Gesa Hartwigsen
Luyao Chen

Hierarchical syntactic structure processing is proposed to be at the core of the human language faculty. Syntactic processing is supported by the left fronto-temporal language network, including a core area in the inferior frontal gyrus as well as its interaction with the posterior temporal lobe (i. e. , “IFG + pTL”). Moreover, during complex syntactic processes, left IFG also interacts with executive control regions, such as the superior parietal lobule (SPL). However, the functional relevance of these network interactions is largely unclear. In particular, it remains to be demonstrated whether the language network plays a specific causal role in comparatively challenging syntactic processes, separable from the interaction between IFG and other general cognitive regions (i. e. , “IFG + SPL” in the present study). The present study was designed to address this question. Thirty healthy adult Chinese native speakers underwent four continuous theta-burst stimulation (cTBS) sessions: stimulation over IFG, stimulation over IFG + pTL, stimulation over IFG + SPL, and sham stimulation over IFG + irrelevant region in a pseudo-randomized order. In each session, participants were required to label the syntactic categories of jabberwocky sequences retaining real Chinese function words (e. g. , “ム了ウ” is labeled as a verb phrase (VP): “[VP [V了]N]”, similar to “ziff-ed a wug”, where “ziff” and “wug” are nonsense pseudowords, and the whole phrase is a VP). Contrasted with sham cTBS, change percentage of accuracy rates (ΔACCR%), reaction times (ΔRT%), and coefficient of variation (ΔCV%) were calculated and compared across conditions. First-order behavioral results showed a significantly higher ΔCV% after stimulating IFG + pTL compared to stimulating the IFG + SPL, indicating that syntactic processing became more unstable. Second-order representational similarity analysis (RSA) results revealed that cTBS effects on IFG + pTL selectively depended on the hierarchical embedding depth, a key measure of syntactic hierarchical complexity, whereas the effects on IFG + SPL were sensitive to the dependency length, a crucial index reflecting the working memory load. Collectively, these findings reveal the specific causal relevance of the language areas for hierarchical syntactic processing, separable from other general cognitive (such as working memory) capacities. These results shed light on the uniqueness and the specific causal role of the language network for the human language faculty, further supporting the causally separable view of the functional dissociation between the language network and the domain-general/multiple-demand network.

Details DOI

ICML Conference 2024 Conference Paper

InfiAgent-DABench: Evaluating Agents on Data Analysis Tasks

Xueyu Hu
Ziyu Zhao 0001
Shuang Wei
Ziwei Chai
Qianli Ma
Guoyin Wang 0002
Xuwu Wang
Jing Su

In this paper, we introduce InfiAgent-DABench, the first benchmark specifically designed to evaluate LLM-based agents on data analysis tasks. Agents need to solve these tasks end-to-end by interacting with an execution environment. This benchmark contains DAEval, a dataset consisting of 603 data analysis questions derived from 124 CSV files, and an agent framework which incorporates LLMs to serve as data analysis agents for both serving and evaluating. Since data analysis questions are often open-ended and hard to evaluate without human supervision, we adopt a format-prompting technique to convert each question into a closed-form format so that they can be automatically evaluated. Our extensive benchmarking of 34 LLMs uncovers the current challenges encountered in data analysis tasks. In addition, building upon our agent framework, we develop a specialized agent, DAAgent, which surpasses GPT-3. 5 by 3. 9% on DABench. Evaluation datasets and toolkits for InfiAgent-DABench are released at https: //github. com/InfiAgent/InfiAgent.

Details

EAAI Journal 2024 Journal Article

Optimized single-image super-resolution reconstruction: A multimodal approach based on reversible guidance and cyclical knowledge distillation

JingKe Yan
Qin Wang
Yao Cheng
ZhaoYu Su
Fan Zhang
MeiLing Zhong
Lei Liu
Bo Jin

This paper proposes a new approach for reconstructing high-resolution images from low-resolution inputs using Denoising Diffusion Probabilistic Models (DDPMs). Existing DDPMs, while promising, face two issues: one is detail discrepancies due to the uncertain degradation factors in low-resolution images, the other is slow sampling speeds. To address these, a multimodal approach based on reversible guidance and cyclical knowledge distillation (MRKD) is introduced. This method is based on the concept where prior and posterior probabilities can assist in comprehending and predicting future events from available data and information. In the MRKD method, text and image information are separately encoded, and novel constraints are applied on prior and posterior distributions, optimizing the detailed features of the reconstructed image. In addition, due to the uncertainty of degradation factors in low-resolution images, a ‘one-to-many’ mapping issue arises in single-image super-resolution tasks. In response to this, the paper redefines constraints on the posterior distribution using the log-likelihood. Specifically, the Bayesian transformation of the input and output of the observation model is employed to effectively guide the diffusion process. To boost the slow sampling speed of DDPM, a cyclical knowledge distillation strategy is proposed, allowing iterative transfer of learned parameters from a high-step DDPM to a low-step model, thereby accelerating the sampling process while preserving image quality. The experimental results demonstrate that these strategies enable the model to effectively comprehend the high-level semantics and contextual information within images. Additionally, they address challenges associated with mode collapse, the loss of high-frequency details, and the complexities of long-tail data.

Details DOI