arXiv subjects

Huan Liu

Publications and source records attributed to Huan Liu.

At least 127 records · Page 7Linked to original sources

arXiv2025

Are We Merely Justifying Results ex Post Facto? Quantifying Explanatory Inversion in Post-Hoc Model Explanations

Post-hoc explanation methods provide interpretation by attributing predictions to input features. Natural explanations are expected to interpret how the inputs lead to the predictions. Thus, a fundamental question arises: Do these explanations unintentionally reverse the natural relationship between inputs and outputs? Specifically, are the explanations rationalizing predictions from the output rather than reflecting the true decision process? To investigate such explanatory inversion, we propose Inversion Quantification (IQ), a framework that quantifies the degree to which explanations rely on outputs and deviate from faithful input-output relationships. Using the framework, we demonstrate on synthetic datasets that widely used methods such as LIME and SHAP are prone to such inversion, particularly in the presence of spurious correlations, across tabular, image, and text domains. Finally, we propose Reproduce-by-Poking (RBP), a simple and model-agnostic enhancement to post-hoc explanation methods that integrates forward perturbation checks. We further show that under the IQ framework, RBP theoretically guarantees the mitigation of explanatory inversion. Empirically, for example, on the synthesized data, RBP can reduce the inversion by 1.8% on average across iconic post-hoc explanation approaches and domains.

cs.LG↗

arXiv2025

Window Token Concatenation for Efficient Visual Large Language Models

To effectively reduce the visual tokens in Visual Large Language Models (VLLMs), we propose a novel approach called Window Token Concatenation (WiCo). Specifically, we employ a sliding window to concatenate spatially adjacent visual tokens. However, directly concatenating these tokens may group diverse tokens into one, and thus obscure some fine details. To address this challenge, we propose fine-tuning the last few layers of the vision encoder to adaptively adjust the visual tokens, encouraging that those within the same window exhibit similar features. To further enhance the performance on fine-grained visual understanding tasks, we introduce WiCo+, which decomposes the visual tokens in later layers of the LLM. Such a design enjoys the merits of the large perception field of the LLM for fine-grained visual understanding while keeping a small number of visual tokens for efficient inference. We perform extensive experiments on both coarse- and fine-grained visual understanding tasks based on LLaVA-1.5 and Shikra, showing better performance compared with existing token reduction projectors. The code is available: https://github.com/JackYFL/WiCo.

cs.CV↗

arXiv2025

Beyond Accuracy: The Role of Calibration in Self-Improving Large Language Models

Large Language Models (LLMs) have demonstrated remarkable self-improvement capabilities, whereby models iteratively revise their outputs through self-generated feedback. While this reflective mechanism has shown promise in enhancing task performance, recent studies suggest that it may also introduce undesirable biases-most notably, self-bias, or the tendency of LLMs to favor their own prior outputs. In this work, we extend this line of inquiry by investigating the impact on confidence estimation. We evaluate three representative self-improvement paradigms-basic prompting, Chain-of-Thought (CoT) prompting, and tuning-based methods and find that iterative self-improvement can lead to systematic overconfidence, as evidenced by a steadily increasing Expected Calibration Error (ECE) and lower accuracy with high confidence. We then further explore the integration of confidence calibration techniques with self-improvement. Specifically, we compare three strategies: (1) applying calibration after multiple rounds of self-improvement, (2) calibrating before self-improvement, and (3) applying calibration iteratively at each self-improvement step. Our results show that iterative calibration is most effective in reducing ECE, yielding improved calibration. Our work pioneers the study of self-improving LLMs from a calibration perspective, offering valuable insights into balancing model performance and reliability.

cs.CL↗

arXiv2025

Navigating Decentralized Online Social Networks: An Overview of Technical and Societal Challenges in Architectural Choices

Decentralized online social networks have evolved from experimental stages to operating at unprecedented scale, with broader adoption and more active use than ever before. Platforms like Mastodon, Bluesky, Hive, and Nostr have seen notable growth, particularly following the wave of user migration after Twitter's acquisition in October 2022. As new platforms build upon earlier decentralization architectures and explore novel configurations, it becomes increasingly important to understand how these foundations shape both the direction and limitations of decentralization. Prior literature primarily focuses on specific architectures, resulting in fragmented views that overlook how different social networks encounter similar challenges and complement one another. This paper fills that gap by presenting a comprehensive view of the current decentralized online social network landscape. We examine four major architectures: federated, peer-to-peer, blockchain, and hybrid, tracing their evolution and evaluating how they support core social networking functions. By linking these architectural aspects to real-world cases, our work provides a foundation for understanding the societal implications of decentralized social platforms.

cs.SI↗

arXiv2025

RedditESS: A Mental Health Social Support Interaction Dataset -- Understanding Effective Social Support to Refine AI-Driven Support Tools

Effective mental health support is crucial for alleviating psychological distress. While large language model (LLM)-based assistants have shown promise in mental health interventions, existing research often defines "effective" support primarily in terms of empathetic acknowledgments, overlooking other essential dimensions such as informational guidance, community validation, and tangible coping strategies. To address this limitation and better understand what constitutes effective support, we introduce RedditESS, a novel real-world dataset derived from Reddit posts, including supportive comments and original posters' follow-up responses. Grounded in established social science theories, we develop an ensemble labeling mechanism to annotate supportive comments as effective or not and perform qualitative assessments to ensure the reliability of the annotations. Additionally, we demonstrate the practical utility of RedditESS by using it to guide LLM alignment toward generating more context-sensitive and genuinely helpful supportive responses. By broadening the understanding of effective support, our study paves the way for advanced AI-driven mental health interventions.

cs.CL↗

arXiv2025

Label Distribution Learning-Enhanced Dual-KNN for Text Classification

Many text classification methods usually introduce external information (e.g., label descriptions and knowledge bases) to improve the classification performance. Compared to external information, some internal information generated by the model itself during training, like text embeddings and predicted label probability distributions, are exploited poorly when predicting the outcomes of some texts. In this paper, we focus on leveraging this internal information, proposing a dual $k$ nearest neighbor (D$k$NN) framework with two $k$NN modules, to retrieve several neighbors from the training set and augment the distribution of labels. For the $k$NN module, it is easily confused and may cause incorrect predictions when retrieving some nearest neighbors from noisy datasets (datasets with labeling errors) or similar datasets (datasets with similar labels). To address this issue, we also introduce a label distribution learning module that can learn label similarity, and generate a better label distribution to help models distinguish texts more effectively. This module eases model overfitting and improves final classification performance, hence enhancing the quality of the retrieved neighbors by $k$NN modules during inference. Extensive experiments on the benchmark datasets verify the effectiveness of our method.

cs.CL↗

arXiv2025

Constrained Gaussian Wasserstein Optimal Transport with Commutative Covariance Matrices

Optimal transport has found widespread applications in signal processing and machine learning. Among its many equivalent formulations, optimal transport seeks to reconstruct a random variable/vector with a prescribed distribution at the destination while minimizing the expected distortion relative to a given random variable/vector at the source. However, in practice, certain constraints may render the optimal transport plan infeasible. In this work, we consider three types of constraints: rate constraints, dimension constraints, and channel constraints, motivated by perception-aware lossy compression, generative principal component analysis, and deep joint source-channel coding, respectively. Special attenion is given to the setting termed Gaussian Wasserstein optimal transport, where both the source and reconstruction variables are multivariate Gaussian, and the end-to-end distortion is measured by the mean squared error. We derive explicit results for the minimum achievable mean squared error under the three aforementioned constraints when the covariance matrices of the source and reconstruction variables commute.

cs.IT↗

arXiv2025

Can LLMs Improve Multimodal Fact-Checking by Asking Relevant Questions?

Traditional fact-checking relies on humans to formulate relevant and targeted fact-checking questions (FCQs), search for evidence, and verify the factuality of claims. While Large Language Models (LLMs) have been commonly used to automate evidence retrieval and factuality verification at scale, their effectiveness for fact-checking is hindered by the absence of FCQ formulation. To bridge this gap, we seek to answer two research questions: (1) Can LLMs generate relevant FCQs? (2) Can LLM-generated FCQs improve multimodal fact-checking? We therefore introduce a framework LRQ-FACT for using LLMs to generate relevant FCQs to facilitate evidence retrieval and enhance fact-checking by probing information across multiple modalities. Through extensive experiments, we verify if LRQ-FACT can generate relevant FCQs of different types and if LRQ-FACT can consistently outperform baseline methods in multimodal fact-checking. Further analysis illustrates how each component in LRQ-FACT works toward improving the fact-checking performance.

cs.CL↗

arXiv2025

Comprehensive scaling laws across animals, microorganisms and plants

Scaling laws illuminate Nature's fundamental biological principles and guide bioinspired materials and structural designs. In simple cases they are based on the fundamental principle that all laws of nature remain unchanged (i.e., invariant) under a change of units. A more general framework is a change of variables for the governing laws that takes all equations, boundary, and interaction conditions into themselves. We consider an accepted macroscale system of partial differential equations including coupled fluid dynamics, nonlinear elasticity, and rigid body mechanics for a complex organism. We show that there is a set of scaling laws where length, time, density, elastic modulus, viscosity, and gravitational constant undergo nontrivial scaling (Table 1). We compare these results to extensive data sets mined from the literature on beating frequency of flying, swimming, and running animals, speed of bacteria, insects, fish, mammals and reptiles, leg stiffness of mammals, and modulus of elasticity of plants. The uniform agreement of the scaling laws with the dynamics of fauna, flora, and microorganisms supports the dominating role of coupled nonlinear elasticity and fluid dynamics in evolutionary development. We conclude with predictions for some prehistoric cases for which observations are unavailable.

physics.bio-ph↗

arXiv2025

MIND-EEG: Multi-granularity Integration Network with Discrete Codebook for EEG-based Emotion Recognition

Emotion recognition using electroencephalogram (EEG) signals has broad potential across various domains. EEG signals have ability to capture rich spatial information related to brain activity, yet effectively modeling and utilizing these spatial relationships remains a challenge. Existing methods struggle with simplistic spatial structure modeling, failing to capture complex node interactions, and lack generalizable spatial connection representations, failing to balance the dynamic nature of brain networks with the need for discriminative and generalizable features. To address these challenges, we propose the Multi-granularity Integration Network with Discrete Codebook for EEG-based Emotion Recognition (MIND-EEG). The framework employs a multi-granularity approach, integrating global and regional spatial information through a Global State Encoder, an Intra-Regional Functionality Encoder, and an Inter-Regional Interaction Encoder to comprehensively model brain activity. Additionally, we introduce a discrete codebook mechanism for constructing network structures via vector quantization, ensuring compact and meaningful brain network representations while mitigating over-smoothing and enhancing model generalization. The proposed framework effectively captures the dynamic and diverse nature of EEG signals, enabling robust emotion recognition. Extensive comparisons and analyses demonstrate the effectiveness of MIND-EEG, and the source code is publicly available at https://anonymous.4open.science/r/MIND_EEG.

cs.HC↗

arXiv2025

A machine-learning optimized vertical-axis wind turbine

Vertical-axis wind turbines (VAWTs) have garnered increasing attention in the field of renewable energy due to their unique advantages over traditional horizontal-axis wind turbines (HAWTs). However, traditional VAWTs including Darrieus and Savonius types suffer from significant drawbacks -- negative torque regions exist during rotation. In this work, we propose a new design of VAWT, which combines design principles from both Darrieus and Savonius but addresses their inherent defects. The performance of the proposed VAWT is evaluated through numerical simulations and validated by experimental testing. The results demonstrate that its power output is approximately three times greater than that of traditional Savonius VAWTs of comparable size. The performance of the proposed VAWT is further optimized using machine learning techniques, including Gaussian process regression and neural networks, based on extensive supercomputer simulations. This optimization leads to a 30% increase in power output.

eess.SP↗

arXiv2025

Sibyl: Empowering Empathetic Dialogue Generation in Large Language Models via Sensible and Visionary Commonsense Inference

Recently, there has been a heightened interest in building chatbots based on Large Language Models (LLMs) to emulate human-like qualities in multi-turn conversations. Despite having access to commonsense knowledge to better understand the psychological aspects and causality of dialogue context, even these powerful LLMs struggle to achieve the goals of empathy and emotional support. Current commonsense knowledge derived from dialogue contexts is inherently limited and often fails to adequately anticipate the future course of a dialogue. This lack of foresight can mislead LLMs and hinder their ability to provide effective support. In response to this challenge, we present an innovative framework named Sensible and Visionary Commonsense Knowledge (Sibyl). Designed to concentrate on the immediately succeeding dialogue, this paradigm equips LLMs with the capability to uncover the implicit requirements of the conversation, aiming to elicit more empathetic responses. Experimental results demonstrate that incorporating our paradigm for acquiring commonsense knowledge into LLMs comprehensively enhances the quality of their responses.

cs.CL↗

arXiv2025

A Multi-Label EEG Dataset for Mental Attention State Classification in Online Learning

Attention is a vital cognitive process in the learning and memory environment, particularly in the context of online learning. Traditional methods for classifying attention states of online learners based on behavioral signals are prone to distortion, leading to increased interest in using electroencephalography (EEG) signals for authentic and accurate assessment. However, the field of attention state classification based on EEG signals in online learning faces challenges, including the scarcity of publicly available datasets, the lack of standardized data collection paradigms, and the requirement to consider the interplay between attention and other psychological states. In light of this, we present the Multi-label EEG dataset for classifying Mental Attention states (MEMA) in online learning. We meticulously designed a reliable and standard experimental paradigm with three attention states: neutral, relaxing, and concentrating, considering human physiological and psychological characteristics. This paradigm collected EEG signals from 20 subjects, each participating in 12 trials, resulting in 1,060 minutes of data. Emotional state labels, basic personal information, and personality traits were also collected to investigate the relationship between attention and other psychological states. Extensive quantitative and qualitative analysis, including a multi-label correlation study, validated the quality of the EEG attention data. The MEMA dataset and analysis provide valuable insights for advancing research on attention in online learning. The dataset is publicly available at \url{https://github.com/GuanjianLiu/MEMA}.

cs.HC↗

arXiv2025

Riemann--Hilbert method to the Ablowitz--Ladik equation: higher-order case

We focused on the Ablowitz--Ladik equation on a zero background, specifically considering the scenario of $N$ pairs of multiple poles. Our first goal was to establish a mapping between the initial data and the scattering data. This allowed us to introduce a direct problem by analyzing the discrete spectrum associated with $N$ pairs of higher-order zeros. Next, we constructed another mapping from the scattering data to a $2\times2$ matrix Riemann--Hilbert problem equipped with several residue conditions set at $N$ pairs of multiple poles. By characterizing the inverse problem based on this Riemann--Hilbert problem, we were able to derive higher-order soliton solutions in the reflectionless case. Furthermore, we expressed an infinite-order soliton solution using a special Riemann--Hilbert problem formulation.

nlin.SI↗

arXiv2025

Real-Valued Vector Modified Korteweg--de Vries Equation: Solitons Featuring Multiple Poles

We delve into the inverse scattering transform of the real-valued vector modified Korteweg--de Vries equation, emphasizing the challenges posed by $N$ pairs of higher-order poles in the transmission coefficient and the enhanced spectral symmetry stemming from real-valued constraints. Utilizing the generalized vector cross product, we formulate an $(n+1)\times(n+1)$ matrix-valued Riemann--Hilbert problem to tackle the complexities inherent in multi-component systems. We subsequently demonstrate the existence and uniqueness of solutions for a singularity-free equivalent problem, adeptly handling the intricacies of multiple poles. In reflectionless cases, we reconstruct multi-pole soliton solutions through a system of linear algebraic equations.

nlin.PS↗

arXiv2025

Visual Large Language Models for Generalized and Specialized Applications

Visual-language models (VLM) have emerged as a powerful tool for learning a unified embedding space for vision and language. Inspired by large language models, which have demonstrated strong reasoning and multi-task capabilities, visual large language models (VLLMs) are gaining increasing attention for building general-purpose VLMs. Despite the significant progress made in VLLMs, the related literature remains limited, particularly from a comprehensive application perspective, encompassing generalized and specialized applications across vision (image, video, depth), action, and language modalities. In this survey, we focus on the diverse applications of VLLMs, examining their using scenarios, identifying ethics consideration and challenges, and discussing future directions for their development. By synthesizing these contents, we aim to provide a comprehensive guide that will pave the way for future innovations and broader applications of VLLMs. The paper list repository is available: https://github.com/JackYFL/awesome-VLLMs.

cs.CV↗

arXiv2024

Revisiting MLLMs: An In-Depth Analysis of Image Classification Abilities

With the rapid advancement of Multimodal Large Language Models (MLLMs), a variety of benchmarks have been introduced to evaluate their capabilities. While most evaluations have focused on complex tasks such as scientific comprehension and visual reasoning, little attention has been given to assessing their fundamental image classification abilities. In this paper, we address this gap by thoroughly revisiting the MLLMs with an in-depth analysis of image classification. Specifically, building on established datasets, we examine a broad spectrum of scenarios, from general classification tasks (e.g., ImageNet, ObjectNet) to more fine-grained categories such as bird and food classification. Our findings reveal that the most recent MLLMs can match or even outperform CLIP-style vision-language models on several datasets, challenging the previous assumption that MLLMs are bad at image classification \cite{VLMClassifier}. To understand the factors driving this improvement, we conduct an in-depth analysis of the network architecture, data selection, and training recipe used in public MLLMs. Our results attribute this success to advancements in language models and the diversity of training data sources. Based on these observations, we further analyze and attribute the potential reasons to conceptual knowledge transfer and enhanced exposure of target concepts, respectively. We hope our findings will offer valuable insights for future research on MLLMs and their evaluation in image classification tasks.

cs.CV↗

arXiv2024

Band Offsets at β/γ-$\mathrm{Ga}_{2}\mathrm{O}_{3}$ Interface

Ultrawide bandgap semiconductor gallium oxide (Ga2O3) and its polymorphs have recently attracted increasing attention across physics, materials science, and electronics communities. In particular, the self-organized formation of the beta/gamma-Ga2O3 double polymorph structures was demonstrated recently [A. Azarov et al., Nat. Commun. 14, 4855 (2023)], paving the way for prospective applications of such structures in electronics. Consequently, determining the conduction band offset in such structures is crucial since it dictates the behavior of conduction electrons at the interface and, consequently, the potential functionality of such interfaces. Thus, in this work, we calculate the band offsets at the beta/gamma-Ga2O3 interface using density functional theory in correlation with the data provided by the experimental atomistic interface analysis. Specifically, to unravel the strain state of the beta/gamma-Ga2O3 interface, nanoscale strain maps were recorded using high-resolution transmission electron microscopy. In its turn, theoretically, lineup potential and vacuum alignment methods were used to analyze the band offsets, with and without strain, at the beta/gamma-Ga2O3 interface. Altogether, the collected results suggest that the band offsets between the beta and gamma phases are likely not exceeding a few hundred meV, remaining highly sensitive to the strain state at the interface. At this end, we conclude that even though the formation of a two-dimensional electron gas (2DEG) at the beta/gamma interface is theoretically possible, the gradual strain relaxation--if it occurs as a function of the distance from the interface--poses a significant challenge, as it may shift the 2DEG localization or even reduce the overall probability of its formation.

cond-mat.mtrl-sci↗