arXiv subjects

Chenhui Qiang

Publications and source records attributed to Chenhui Qiang.

4 recordsLinked to original sources

arXiv2025

VER-Bench: Evaluating MLLMs on Reasoning with Fine-Grained Visual Evidence

With the rapid development of MLLMs, evaluating their visual capabilities has become increasingly crucial. Current benchmarks primarily fall into two main types: basic perception benchmarks, which focus on local details but lack deep reasoning (e.g., "what is in the image?"), and mainstream reasoning benchmarks, which concentrate on prominent image elements but may fail to assess subtle clues requiring intricate analysis. However, profound visual understanding and complex reasoning depend more on interpreting subtle, inconspicuous local details than on perceiving salient, macro-level objects. These details, though occupying minimal image area, often contain richer, more critical information for robust analysis. To bridge this gap, we introduce the VER-Bench, a novel framework to evaluate MLLMs' ability to: 1) identify fine-grained visual clues, often occupying on average just 0.25% of the image area; 2) integrate these clues with world knowledge for complex reasoning. Comprising 374 carefully designed questions across Geospatial, Temporal, Situational, Intent, System State, and Symbolic reasoning, each question in VER-Bench is accompanied by structured evidence: visual clues and question-related reasoning derived from them. VER-Bench reveals current models' limitations in extracting subtle visual evidence and constructing evidence-based arguments, highlighting the need to enhance models's capabilities in fine-grained visual evidence extraction, integration, and reasoning for genuine visual understanding and human-like analysis. Dataset and additional materials are available https://github.com/verbta/ACMMM-25-Materials.

cs.CV↗

arXiv2024

ViMoE: An Empirical Study of Designing Vision Mixture-of-Experts

Mixture-of-Experts (MoE) models embody the divide-and-conquer concept and are a promising approach for increasing model capacity, demonstrating excellent scalability across multiple domains. In this paper, we integrate the MoE structure into the classic Vision Transformer (ViT), naming it ViMoE, and explore the potential of applying MoE to vision through a comprehensive study on image classification and semantic segmentation. However, we observe that the performance is sensitive to the configuration of MoE layers, making it challenging to obtain optimal results without careful design. The underlying cause is that inappropriate MoE layers lead to unreliable routing and hinder experts from effectively acquiring helpful information. To address this, we introduce a shared expert to learn and capture common knowledge, serving as an effective way to construct stable ViMoE. Furthermore, we demonstrate how to analyze expert routing behavior, revealing which MoE layers are capable of specializing in handling specific information and which are not. This provides guidance for retaining the critical layers while removing redundancies, thereby advancing ViMoE to be more efficient without sacrificing accuracy. We aspire for this work to offer new insights into the design of vision MoE models and provide valuable empirical guidance for future research.

cs.CV↗

arXiv2021

Information fractal dimension of mass function

Fractal plays an important role in nonlinear science. The most important parameter to model fractal is fractal dimension. Existing information dimension can calculate the dimension of probability distribution. However, given a mass function which is the generalization of probability distribution, how to determine its fractal dimension is still an open problem of immense interest. The main contribution of this work is to propose an information fractal dimension of mass function. Numerical examples are illustrated to show the effectiveness of our proposed dimension. We discover an important property in that the dimension of mass function with the maximum Deng entropy is $\frac{ln3}{ln2}\approx 1.585$, which is the well-known fractal dimension of Sierpiński triangle.

cs.IT↗

arXiv2021

Multifractal of mass function

Multifractal plays an important role in many fields. However, there is few attentions about mass function, which can better deal with uncertain information than probability. In this paper, we proposed multifractal of mass function. Firstly, the definition of multifractal spectrum of mass function is given. Secondly, the multifractal dimension of mass function is defined as $D_α$. When mass function degenerates to probability distribution, $D_α$ degenerates to $d_α$, which is information dimension proposed by Renyi. One interesting property is that the multifractal dimension of mass function with maximum Deng entropy is 1.585 no matter the order. Other interesting properties and numerical examples are shown to illustrate proposed model.

cs.IT↗