Search Results for author: Xinyu Zhao

Found 32 papers, 9 papers with code

ChartMuseum: Testing Visual Reasoning Capabilities of Large Vision-Language Models

no code implementations19 May 2025 Liyan Tang, Grace Kim, Xinyu Zhao, Thom Lake, Wenxuan Ding, Fangcong Yin, Prasann Singhal, Manya Wadhwa, Zeyu Leo Liu, Zayne Sprague, Ramya Namuduri, Bodun Hu, Juan Diego Rodriguez, Puyuan Peng, Greg Durrett

Unlike prior chart understanding benchmarks -- where frontier models perform similarly and near saturation -- our benchmark exposes a substantial gap between model and human performance, while effectively differentiating model capabilities: although humans achieve 93% accuracy, the best-performing model Gemini-2. 5-Pro attains only 63. 0%, and the leading open-source LVLM Qwen2. 5-VL-72B-Instruct achieves only 38. 5%.

Chart Question Answering Chart Understanding +2

Contour Field based Elliptical Shape Prior for the Segment Anything Model

no code implementations17 Apr 2025 Xinyu Zhao, Jun Liu, Faqiang Wang, Li Cui, Yuping Duan

This paper proposes a new approach to integrate the prior of elliptical shapes into the deep learning-based SAM image segmentation techniques using variational methods.

Image Segmentation Segmentation +1

SAFE: Self-Adjustment Federated Learning Framework for Remote Sensing Collaborative Perception

no code implementations25 Mar 2025 Xiaohe Li, Haohua Wu, Jiahao Li, Zide Fan, Kaixin Zhang, Xinming Li, Yunping Ge, Xinyu Zhao

Experiments on real-world image classification and object segmentation datasets validate the effectiveness and reliability of the SAFE framework in complex remote sensing scenarios.

Computational Efficiency Federated Learning +3

Dialogue is Better Than Monologue: Instructing Medical LLMs via Strategical Conversations

no code implementations29 Jan 2025 Zijie Liu, Xinyu Zhao, Jie Peng, Zhuangdi Zhu, Qingyu Chen, Xia Hu, Tianlong Chen

Current medical AI systems often fail to replicate real-world clinical reasoning, as they are predominantly trained and evaluated on static text and question-answer tasks.

Diagnostic

Convex Combination Star Shape Prior for Data-driven Image Semantic Segmentation

no code implementations CVPR 2025 Xinyu Zhao, Jun Xie, Shengzhe Chen, Jun Liu

Multi-center star shape is a prevalent object shape feature, which has proven effective in model-based image segmentation methods.

Image Segmentation Segmentation +1

BrainMAP: Learning Multiple Activation Pathways in Brain Networks

2 code implementations23 Dec 2024 Song Wang, Zhenyu Lei, Zhen Tan, Jiaqi Ding, Xinyu Zhao, Yushun Dong, Guorong Wu, Tianlong Chen, Chen Chen, Aiying Zhang, Jundong Li

As such, conventional GNNs struggle to learn from these pathways due to the long-range dependencies of multiple pathways.

Mamba Mixture-of-Experts

Towards Stabilized and Efficient Diffusion Transformers through Long-Skip-Connections with Spectral Constraints

1 code implementation26 Nov 2024 Guanjie Chen, Xinyu Zhao, Yucheng Zhou, Xiaoye Qu, Tianlong Chen, Yu Cheng

Diffusion Transformers (DiT) have emerged as a powerful architecture for image and video generation, offering superior quality and scalability.

Denoising Image Generation +1

Understanding Synthetic Context Extension via Retrieval Heads

no code implementations29 Oct 2024 Xinyu Zhao, Fangcong Yin, Greg Durrett

To defray the cost of pretraining LLMs over long contexts, recent work takes an approach of synthetic context extension: fine-tuning LLMs with synthetically generated long-context data in a post-training stage.

Retrieval Retrieval-augmented Generation

Model-GLUE: Democratized LLM Scaling for A Large Model Zoo in the Wild

1 code implementation7 Oct 2024 Xinyu Zhao, Guoheng Sun, Ruisi Cai, Yukun Zhou, Pingzhi Li, Peihao Wang, Bowen Tan, Yexiao He, Li Chen, Yi Liang, Beidi Chen, Binhang Yuan, Hongyi Wang, Ang Li, Zhangyang Wang, Tianlong Chen

As Large Language Models (LLMs) excel across tasks and specialized domains, scaling LLMs based on existing models has garnered significant attention, which faces the challenge of decreasing performance when combining disparate models.

Benchmarking Mixture-of-Experts +1

To CoT or not to CoT? Chain-of-thought helps mainly on math and symbolic reasoning

1 code implementation18 Sep 2024 Zayne Sprague, Fangcong Yin, Juan Diego Rodriguez, Dongwei Jiang, Manya Wadhwa, Prasann Singhal, Xinyu Zhao, Xi Ye, Kyle Mahowald, Greg Durrett

Chain-of-thought (CoT) via prompting is the de facto method for eliciting reasoning capabilities from large language models (LLMs).

Math MMLU

SDoH-GPT: Using Large Language Models to Extract Social Determinants of Health (SDoH)

no code implementations24 Jul 2024 Bernardo Consoli, Xizhi Wu, Song Wang, Xinyu Zhao, Yanshan Wang, Justin Rousseau, Tom Hartvigsen, Li Shen, Huanmei Wu, Yifan Peng, Qi Long, Tianlong Chen, Ying Ding

Extracting social determinants of health (SDoH) from unstructured medical notes depends heavily on labor-intensive annotations, which are typically task-specific, hampering reusability and limiting sharing.

Computational Efficiency Language Modeling +2

Learning to Refine with Fine-Grained Natural Language Feedback

1 code implementation2 Jul 2024 Manya Wadhwa, Xinyu Zhao, Junyi Jessy Li, Greg Durrett

Recent work has explored the capability of large language models (LLMs) to identify and correct errors in LLM-generated responses.

$\texttt{MoE-RBench}$: Towards Building Reliable Language Models with Sparse Mixture-of-Experts

1 code implementation17 Jun 2024 Guanjie Chen, Xinyu Zhao, Tianlong Chen, Yu Cheng

Motivated by the research gap and counter-intuitive phenomenon, we propose $\texttt{MoE-RBench}$, the first comprehensive assessment of SMoE reliability from three aspects: $\textit{(i)}$ safety and hallucination, $\textit{(ii)}$ resilience to adversarial attacks, and $\textit{(iii)}$ out-of-distribution robustness.

Hallucination Mixture-of-Experts

GraphRCG: Self-Conditioned Graph Generation

no code implementations2 Mar 2024 Song Wang, Zhen Tan, Xinyu Zhao, Tianlong Chen, Huan Liu, Jundong Li

In contrast, in this work, we propose a novel self-conditioned graph generation framework designed to explicitly model graph distributions and employ these distributions to guide the generation process.

Graph Generation

Enhancing Essay Scoring with Adversarial Weights Perturbation and Metric-specific AttentionPooling

no code implementations6 Jan 2024 Jiaxin Huang, Xinyu Zhao, Chang Che, Qunwei Lin, Bo Liu

To address the specific needs of ELLs, we propose the use of DeBERTa, a state-of-the-art neural language model, for improving automated feedback tools.

Automated Essay Scoring Language Modelling +2

Semantic Similarity Matching for Patent Documents Using Ensemble BERT-related Model and Novel Text Processing Method

no code implementations6 Jan 2024 Liqiang Yu, Bo Liu, Qunwei Lin, Xinyu Zhao, Chang Che

In the realm of patent document analysis, assessing semantic similarity between phrases presents a significant challenge, notably amplifying the inherent complexities of Cooperative Patent Classification (CPC) research.

Patent classification Semantic Similarity +1

Integration and Performance Analysis of Artificial Intelligence and Computer Vision Based on Deep Learning Algorithms

no code implementations20 Dec 2023 Bo Liu, Liqiang Yu, Chang Che, Qunwei Lin, Hao Hu, Xinyu Zhao

This paper focuses on the analysis of the application effectiveness of the integration of deep learning and computer vision technologies.

Deep Learning image-classification +3

CDIDN: A Registration Model with High Deformation Impedance Capability for Long-Term Tracking of Pulmonary Lesion Dynamics

no code implementations18 May 2023 Xinyu Zhao, Sa Huang, Wei Pang, You Zhou

In this paper, we propose a novel registration model called Cascade-Dilation Inter-Layer Differential Network (CDIDN), which exhibits both high deformation impedance capability (DIC) and accuracy.

Efficient Domain Coverage for Vehicles with Second-Order Dynamics via Multi-Agent Reinforcement Learning

no code implementations11 Nov 2022 Xinyu Zhao, Razvan C. Fetecau, Mo Chen

Our proposed network architecture includes the incorporation of LSTM and self-attention, which allows the trained policy to adapt to a variable number of agents.

Multi-agent Reinforcement Learning Reinforcement Learning (RL)

Contrastive Psudo-supervised Classification for Intra-Pulse Modulation of Radar Emitter Signals Using data augmentation

no code implementations13 Oct 2022 HanCong Feng, XinHai Yan, Kaili Jiang, Xinyu Zhao, Bin Tang

The automatic classification of radar waveform is a fundamental technique in electronic countermeasures (ECM). Recent supervised deep learning-based methods have achieved great success in a such classification task. However, those methods require enough labeled samples to work properly and in many circumstances, it is not available. To tackle this problem, in this paper, we propose a three-stages deep radar waveform clustering(DRSC) technique to automatically group the received signal samples without labels. Firstly, a pretext model is trained in a self-supervised way with the help of several data augmentation techniques to extract the class-dependent features. Next, the pseudo-supervised contrastive training is involved to further promote the separation between the extracted class-dependent features. And finally, the unsupervised problem is converted to a semi-supervised classification problem via pseudo label generation.

Classification Clustering +3

Adaptive Partially-Observed Sequential Change Detection and Isolation

no code implementations9 Aug 2022 Xinyu Zhao, Jiuyun Hu, Yajun Mei, Hao Yan

High-dimensional data has become popular due to the easy accessibility of sensors in modern industrial applications.

Change Detection Change Point Detection

Deep Spatio-temporal Sparse Decomposition for Trend Prediction and Anomaly Detection in Cardiac Electrical Conduction

no code implementations20 Sep 2021 Xinyu Zhao, Hao Yan, Zhiyong Hu, Dongping Du

Electrical conduction among cardiac tissue is commonly modeled with partial differential equations, i. e., reaction-diffusion equation, where the reaction term describes cellular stimulation and diffusion term describes electrical propagation.

Anomaly Detection

Flexible Generation of Natural Language Deductions

1 code implementation EMNLP 2021 Kaj Bostrom, Xinyu Zhao, Swarat Chaudhuri, Greg Durrett

Natural language is an attractive representation for this purpose -- it is both highly expressive and easy for humans to understand.

Sentence

Measurement of tunnel coupling in a Si double quantum dot based on charge sensing

no code implementations11 Mar 2021 Xinyu Zhao, Xuedong Hu

In Si quantum dots, valley degree of freedom, in particular the generally small valley splitting and the dot-dependent valley-orbit phase, adds complexities to the low-energy electron dynamics and the associated spin qubit manipulation.

Mesoscale and Nanoscale Physics Quantum Physics

Large eddy simulation/probability density function simulations of the Cambridge turbulent stratified flame series under swirling conditions

no code implementations1 Dec 2020 Hasret Turkeri, Xinyu Zhao, Metin Muradgolu

The parametric studies show that the differential diffusion has a negligible effect on the mean and r. m. s.

Fluid Dynamics

Effective Distant Supervision for Temporal Relation Extraction

2 code implementations EACL (AdaptNLP) 2021 Xinyu Zhao, Shih-ting Lin, Greg Durrett

A principal barrier to training temporal relation extraction models in new domains is the lack of varied, high quality examples and the challenge of collecting more.

Relation Temporal Relation Extraction

Nutational resonances, transitional precession, and precession-averaged evolution in binary black-hole systems

no code implementations5 May 2017 Xinyu Zhao, Michael Kesden, Davide Gerosa

We use analytic solutions for generic spin precession at 2PN order to derive Fourier series for the total and orbital angular momenta in which each term is a sinusoid with frequency $\Omega - n\omega$ for integer $n$.

General Relativity and Quantum Cosmology High Energy Astrophysical Phenomena

Cannot find the paper you are looking for? You can Submit a new open access paper.