arXiv subjects

Yifeng Hu

Publications and source records attributed to Yifeng Hu.

4 recordsLinked to original sources

arXiv2026

SARE: Sample-wise Adaptive Reasoning for Training-free Fine-grained Visual Recognition

Recent advances in Large Vision-Language Models (LVLMs) have enabled training-free Fine-Grained Visual Recognition (FGVR). However, effectively exploiting LVLMs for FGVR remains challenging due to the inherent visual ambiguity of subordinate-level categories. Existing methods predominantly adopt either retrieval-oriented or reasoning-oriented paradigms to tackle this challenge, but both are constrained by two fundamental limitations:(1) They apply the same inference pipeline to all samples without accounting for uneven recognition difficulty, thereby leading to suboptimal accuracy and efficiency; (2) The lack of mechanisms to consolidate and reuse error-specific experience causes repeated failures on similar challenging cases. To address these limitations, we propose SARE, a Sample-wise Adaptive textbfREasoning framework for training-free FGVR. Specifically, SARE adopts a cascaded design that combines fast candidate retrieval with fine-grained reasoning, invoking the latter only when necessary. In the reasoning process, SARE incorporates a self-reflective experience mechanism that leverages past failures to provide transferable discriminative guidance during inference, without any parameter updates. Extensive experiments across 14 datasets substantiate that SARE achieves state-of-the-art performance while substantially reducing computational overhead.

cs.CV↗

arXiv2025

SELECT: Detecting Label Errors in Real-world Scene Text Data

We introduce SELECT (Scene tExt Label Errors deteCTion), a novel approach that leverages multi-modal training to detect label errors in real-world scene text datasets. Utilizing an image-text encoder and a character-level tokenizer, SELECT addresses the issues of variable-length sequence labels, label sequence misalignment, and character-level errors, outperforming existing methods in accuracy and practical utility. In addition, we introduce Similarity-based Sequence Label Corruption (SSLC), a process that intentionally introduces errors into the training labels to mimic real-world error scenarios during training. SSLC not only can cause a change in the sequence length but also takes into account the visual similarity between characters during corruption. Our method is the first to detect label errors in real-world scene text datasets successfully accounting for variable-length labels. Experimental results demonstrate the effectiveness of SELECT in detecting label errors and improving STR accuracy on real-world text datasets, showcasing its practical utility.

cs.CV↗

arXiv2025

BioMARS: A Multi-Agent Robotic System for Autonomous Biological Experiments

Large language models (LLMs) and vision-language models (VLMs) have the potential to transform biological research by enabling autonomous experimentation. Yet, their application remains constrained by rigid protocol design, limited adaptability to dynamic lab conditions, inadequate error handling, and high operational complexity. Here we introduce BioMARS (Biological Multi-Agent Robotic System), an intelligent platform that integrates LLMs, VLMs, and modular robotics to autonomously design, plan, and execute biological experiments. BioMARS uses a hierarchical architecture: the Biologist Agent synthesizes protocols via retrieval-augmented generation; the Technician Agent translates them into executable robotic pseudo-code; and the Inspector Agent ensures procedural integrity through multimodal perception and anomaly detection. The system autonomously conducts cell passaging and culture tasks, matching or exceeding manual performance in viability, consistency, and morphological integrity. It also supports context-aware optimization, outperforming conventional strategies in differentiating retinal pigment epithelial cells. A web interface enables real-time human-AI collaboration, while a modular backend allows scalable integration with laboratory hardware. These results highlight the feasibility of generalizable, AI-driven laboratory automation and the transformative role of language-based reasoning in biological research.

cs.RO↗

arXiv2017

Defects controlled hole doping and multi-valley transport in SnSe single crystals

SnSe is a promising thermoelectric material with record-breaking figure of merit, \textit{i.e., ZT}. As a semiconductor, optimal electrical dosage is the key challenge to maximize \textit{ZT} in SnSe. However, to date a comprehensive understanding of the electronic structure and most critically, the self-hole doping mechanism in SnSe is still absent. Here, we report the highly anisotropic electronic structure of SnSe investigated by both angle-resolved photoemission spectroscopy and quantum transport, in which a unique "\textit{pudding-mold}" shaped valence band with quasi-linear energy dispersion is revealed. We prove that the electrical doping in SnSe is extrinsically controlled by the formation of SnSe$_{2}$ micro-domains induced by local phase segregation. Using different growth methods and conditions, we have achieved wide tuning of hole doping in SnSe, ranging from intrinsic semiconducting behaviour to typical metal with carrier density of $1.23\times 10^{18}$ cm$^{-3}$ at room temperature. The resulting multi-valley transport in $p$-SnSe is characterized by non-saturating weak localization along the armchair axis, due to strong intervalley scattering enhanced by in-plane ferroelectric dipole field of the puckering lattice. Strikingly, quantum oscillations of magnetoresistance reveal three-dimensional electronic structure with unusual interlayer coupling strength in $p$-SnSe, which is correlated to the interweaving of SnSe individual layers by unique point dislocation defects. Our results suggest that defect engineering may provide versatile routes in improving the thermoelectric performance of the SnSe family.

cond-mat.mtrl-sci↗