arXiv subjects

Liang Bai

Publications and source records attributed to Liang Bai.

At least 19 recordsLinked to original sources

arXiv2026

Stochastic MeanFlow Policies: One-Step Generative Control with Entropic Mirror Descent

Online off-policy reinforcement learning (RL) is shaped by two coupled choices: the policy class and the update rule. Gaussian policies are fast and have tractable entropy, but struggle with multimodal action distributions. Generative policies are more expressive, but often require iterative sampling or lack tractable entropy estimates. On the optimisation side, SAC-style soft policy improvement and mirror descent (MD) can be viewed as minimising different KL divergences: the former moves the policy towards a value-induced Boltzmann distribution, while the latter regularises each update against the previous policy. Combining entropy regularisation with an MD constraint is therefore attractive, as it supports exploration while stabilising policy improvement; however, the resulting target can be multimodal and is poorly matched by unimodal Gaussian policies. We propose Stochastic MeanFlow Policies (SMFP), a one-step generative policy class that maps Gaussian noise to actions through a MeanFlow transformation. This stochastic reparameterisation yields a tractable entropy surrogate and allows MeanFlow policies to be trained within off-policy mirror descent under a unified objective for exploratory yet stable improvement. Across seven MuJoCo benchmarks, SMFP improves over Gaussian and generative baselines while retaining single-step inference efficiency.

cs.LG↗

arXiv2026

Rethinking Generalization in Graph Neural Networks: A Structural Complexity Perspective

Graph neural networks (GNNs) have emerged as a fundamental tool for learning from graph-structured data, achieving strong performance across a wide range of applications. However, understanding their generalization capabilities remains challenging due to the complex structural dependencies inherent in such data. Existing generalization analyses largely follow the classical machine learning paradigm, focusing primarily on model complexity while overlooking the fundamental role of graph structure. Therefore, in this work, we systematically investigate this role by asking: does the graph structure actually influence generalization, and if so, by how much? To answer the first question and validate our intuition, we theoretically prove that incorporating more edges into the prediction process transforms the input representations to be overly accommodating to the output model, thereby inducing overfitting. To address the second question, we formulate a structural complexity measure based on the number of effective edges and derive a Rademacher complexity-based generalization bound. In doing so, we demonstrate that GNN generalization depends explicitly on structural complexity, alongside traditional parameter-dependent factors. Motivated by these theoretical findings, we propose a structural entropy regularization method. This approach controls structural complexity by regulating effective edges to balance underfitting and overfitting, ultimately improving the generalization performance of GNNs.

cs.LG↗

arXiv2026

Dissipation-induced Nonlinear Topological Gear Switching

Nonlinear interaction enables topological phenomena impossible in linear systems. A paradigm is nonlinear Thouless pump, where the transport of solitons can be topologically quantized even when band occupation is nonuniform. Such nonlinear quantization traditionally requires a time-periodic Hamiltonian with static nonlinearity and, much as in the linear case, is inherently independent of pumping speed. Instead, we demonstrate a dissipation-induced topological gear switching, where quantized soliton transport can be switched on and off via the adiabatic pumping speed itself. This phenomenon has no counterpart in prior conservative nonlinear pumps, nor in linear non-Hermitian pumps. Crucially, quantization here no longer requires a time-periodic nonlinear Hamiltonian; it stems from a genuinely non-equilibrium mechanism captured by an effective conservative model whose \textit{nonlinearity varies aperiodically in time}. Remarkably, a quantized nonlinear transport can be induced even when this nonlinear aperiodic driving is such that the system is pumped from the linear to nonlinear regimes. Our results open a route toward nonequilibrium nonlinear topological matter, where topological effect is dynamically reconfigurable via time-varying nonlinearities, with experimental implications for photonic, atomic, or superconducting platforms and beyond.

cond-mat.quant-gas↗

arXiv2026

Detecting Complex-Energy Braiding Topology in a Dissipative Atomic Simulator with Transformer-Based Geometric Tomography

Machine learning (ML) is shaping our exploration of topological matter, whose existence is inherently tied to the geometry of quantum states or energy spectra. In non-Hermitian systems, distinctive spectral geometry can lead to topological braiding of complex-energy bands, yet directly probing this topology-geometry interplay remains challenging. Here, we introduce a Transformer-based ML framework to capture this interplay and experimentally demonstrate it in a dissipative cold-atom simulator. Using a Bose-Einstein condensate, we engineer tunable dissipative two-level systems whose complex eigenenergies form braids. Owing to the density-dependent dissipation, the instantaneous energy braids exhibit topologically distinct structures at short and long times. The Transformer not only accurately predicts topological invariants for diverse energy braids but also, through its self-attention mechanism, autonomously highlights band crossings as the governing underlying geometric feature. Our work paves the way for ML-guided exploration of non-Hermitian topological phases in cold atoms and beyond.

cond-mat.quant-gas↗

arXiv2026

Understanding the Role of Rehearsal Scale in Continual Learning under Varying Model Capacities

Rehearsal is one of the key techniques for mitigating catastrophic forgetting and has been widely adopted in continual learning algorithms due to its simplicity and practicality. However, the theoretical understanding of how rehearsal scale influences learning dynamics remains limited. To address this gap, we formulate rehearsal-based continual learning as a multidimensional effectiveness-driven iterative optimization problem, providing a unified characterization across diverse performance metrics. Within this framework, we derive a closed-form analysis of adaptability, memorability, and generalization from the perspective of rehearsal scale. Our results uncover several intriguing and counterintuitive findings. First, rehearsal can impair model's adaptability, in sharp contrast to its traditionally recognized benefits. Second, increasing the rehearsal scale does not necessarily improve memory retention. When tasks are similar and noise levels are low, the memory error exhibits a diminishing lower bound. Finally, we validate these insights through numerical simulations and extended analyses on deep neural networks across multiple real-world datasets, revealing statistical patterns of rehearsal mechanisms in continual learning.

cs.LG↗

arXiv2026

Exploring the Impact of Parameter Update Magnitude on Forgetting and Generalization of Continual Learning

The magnitude of parameter updates are considered a key factor in continual learning. However, most existing studies focus on designing diverse update strategies, while a theoretical understanding of the underlying mechanisms remains limited. Therefore, we characterize model's forgetting from the perspective of parameter update magnitude and formalize it as knowledge degradation induced by task-specific drift in the parameter space, which has not been fully captured in previous studies due to their assumption of a unified parameter space. By deriving the optimal parameter update magnitude that minimizes forgetting, we unify two representative update paradigms, frozen training and initialized training, within an optimization framework for constrained parameter updates. Our theoretical results further reveals that sequence tasks with small parameter distances exhibit better generalization and less forgetting under frozen training rather than initialized training. These theoretical insights inspire a novel hybrid parameter update strategy that adaptively adjusts update magnitude based on gradient directions. Experiments on deep neural networks demonstrate that this hybrid approach outperforms standard training strategies, providing new theoretical perspectives and practical inspiration for designing efficient and scalable continual learning algorithms.

cs.LG↗

arXiv2026

Bipartite Graph Attention-based Clustering for Large-scale scRNA-seq Data

scRNA-seq clustering is a critical task for analyzing single-cell RNA sequencing (scRNA-seq) data, as it groups cells with similar gene expression profiles. Transformers, as powerful foundational models, have been applied to scRNA-seq clustering. Their self-attention mechanism automatically assigns higher attention weights to cells within the same cluster, enhancing the distinction between clusters. Existing methods for scRNA-seq clustering, such as graph transformer-based models, treat each cell as a token in a sequence. Their computational and space complexities are $\mathcal{O}(n^2)$ with respect to the number of cells, limiting their applicability to large-scale scRNA-seq datasets.To address this challenge, we propose a Bipartite Graph Transformer-based clustering model (BGFormer) for scRNA-seq data. We introduce a set of learnable anchor tokens as shared reference points to represent the entire dataset. A bipartite graph attention mechanism is introduced to learn the similarity between cells and anchor tokens, bringing cells of the same class closer together in the embedding space. BGFormer achieves linear computational complexity with respect to the number of cells, making it scalable to large datasets. Experimental results on multiple large-scale scRNA-seq datasets demonstrate the effectiveness and scalability of BGFormer.

cs.LG↗

arXiv2026

LLM-Guided Diagnostic Evidence Alignment for Medical Vision-Language Pretraining under Limited Pairing

Most existing CLIP-style medical vision--language pretraining methods rely on global or local alignment with substantial paired data. However, global alignment is easily dominated by non-diagnostic information, while local alignment fails to integrate key diagnostic evidence. As a result, learning reliable diagnostic representations becomes difficult, which limits their applicability in medical scenarios with limited paired data. To address this issue, we propose an LLM-Guided Diagnostic Evidence Alignment method (LGDEA), which shifts the pretraining objective toward evidence-level alignment that is more consistent with the medical diagnostic process. Specifically, we leverage LLMs to extract key diagnostic evidence from radiology reports and construct a shared diagnostic evidence space, enabling evidence-aware cross-modal alignment and allowing LGDEA to effectively exploit abundant unpaired medical images and reports, thereby substantially alleviating the reliance on paired data. Extensive experimental results demonstrate that our method achieves consistent and significant improvements on phrase grounding, image--text retrieval, and zero-shot classification, and even rivals pretraining methods that rely on substantial paired data.

cs.CV↗

arXiv2025

One-Step Generative Policies with Q-Learning: A Reformulation of MeanFlow

We introduce a one-step generative policy for offline reinforcement learning that maps noise directly to actions via a residual reformulation of MeanFlow, making it compatible with Q-learning. While one-step Gaussian policies enable fast inference, they struggle to capture complex, multimodal action distributions. Existing flow-based methods improve expressivity but typically rely on distillation and two-stage training when trained with Q-learning. To overcome these limitations, we propose to reformulate MeanFlow to enable direct noise-to-action generation by integrating the velocity field and noise-to-action transformation into a single policy network-eliminating the need for separate velocity estimation. We explore several reformulation variants and identify an effective residual formulation that supports expressive and stable policy learning. Our method offers three key advantages: 1) efficient one-step noise-to-action generation, 2) expressive modelling of multimodal action distributions, and 3) efficient and stable policy learning via Q-learning in a single-stage training setup. Extensive experiments on 73 tasks across the OGBench and D4RL benchmarks demonstrate that our method achieves strong performance in both offline and offline-to-online reinforcement learning settings. Code is available at https://github.com/HiccupRL/MeanFlowQL.

cs.LG↗

arXiv2025

FoCLIP: A Feature-Space Misalignment Framework for CLIP-Based Image Manipulation and Detection

The well-aligned attribute of CLIP-based models enables its effective application like CLIPscore as a widely adopted image quality assessment metric. However, such a CLIP-based metric is vulnerable for its delicate multimodal alignment. In this work, we propose \textbf{FoCLIP}, a feature-space misalignment framework for fooling CLIP-based image quality metric. Based on the stochastic gradient descent technique, FoCLIP integrates three key components to construct fooling examples: feature alignment as the core module to reduce image-text modality gaps, the score distribution balance module and pixel-guard regularization, which collectively optimize multimodal output equilibrium between CLIPscore performance and image quality. Such a design can be engineered to maximize the CLIPscore predictions across diverse input prompts, despite exhibiting either visual unrecognizability or semantic incongruence with the corresponding adversarial prompts from human perceptual perspectives. Experiments on ten artistic masterpiece prompts and ImageNet subsets demonstrate that optimized images can achieve significant improvement in CLIPscore while preserving high visual fidelity. In addition, we found that grayscale conversion induces significant feature degradation in fooling images, exhibiting noticeable CLIPscore reduction while preserving statistical consistency with original images. Inspired by this phenomenon, we propose a color channel sensitivity-driven tampering detection mechanism that achieves 91% accuracy on standard benchmarks. In conclusion, this work establishes a practical pathway for feature misalignment in CLIP-based multimodal systems and the corresponding defense method.

cs.CV↗

arXiv2025

Revisit the Imbalance Optimization in Multi-task Learning: An Experimental Analysis

Multi-task learning (MTL) aims to build general-purpose vision systems by training a single network to perform multiple tasks jointly. While promising, its potential is often hindered by "unbalanced optimization", where task interference leads to subpar performance compared to single-task models. To facilitate research in MTL, this paper presents a systematic experimental analysis to dissect the factors contributing to this persistent problem. Our investigation confirms that the performance of existing optimization methods varies inconsistently across datasets, and advanced architectures still rely on costly grid-searched loss weights. Furthermore, we show that while powerful Vision Foundation Models (VFMs) provide strong initialization, they do not inherently resolve the optimization imbalance, and merely increasing data quantity offers limited benefits. A crucial finding emerges from our analysis: a strong correlation exists between the optimization imbalance and the norm of task-specific gradients. We demonstrate that this insight is directly applicable, showing that a straightforward strategy of scaling task losses according to their gradient norms can achieve performance comparable to that of an extensive and computationally expensive grid search. Our comprehensive analysis suggests that understanding and controlling gradient dynamics is a more direct path to stable MTL than developing increasingly complex methods.

cs.CV↗

arXiv2025

COLA: Context-aware Language-driven Test-time Adaptation

Test-time adaptation (TTA) has gained increasing popularity due to its efficacy in addressing ``distribution shift'' issue while simultaneously protecting data privacy. However, most prior methods assume that a paired source domain model and target domain sharing the same label space coexist, heavily limiting their applicability. In this paper, we investigate a more general source model capable of adaptation to multiple target domains without needing shared labels. This is achieved by using a pre-trained vision-language model (VLM), \egno, CLIP, that can recognize images through matching with class descriptions. While the zero-shot performance of VLMs is impressive, they struggle to effectively capture the distinctive attributes of a target domain. To that end, we propose a novel method -- Context-aware Language-driven TTA (COLA). The proposed method incorporates a lightweight context-aware module that consists of three key components: a task-aware adapter, a context-aware unit, and a residual connection unit for exploring task-specific knowledge, domain-specific knowledge from the VLM and prior knowledge of the VLM, respectively. It is worth noting that the context-aware module can be seamlessly integrated into a frozen VLM, ensuring both minimal effort and parameter efficiency. Additionally, we introduce a Class-Balanced Pseudo-labeling (CBPL) strategy to mitigate the adverse effects caused by class imbalance. We demonstrate the effectiveness of our method not only in TTA scenarios but also in class generalisation tasks. The source code is available at https://github.com/NUDT-Bai-Group/COLA-TTA.

cs.CV↗

arXiv2025

Sculpting Margin Penalty: Intra-Task Adapter Merging and Classifier Calibration for Few-Shot Class-Incremental Learning

Real-world applications often face data privacy constraints and high acquisition costs, making the assumption of sufficient training data in incremental tasks unrealistic and leading to significant performance degradation in class-incremental learning. Forward-compatible learning, which prospectively prepares for future tasks during base task training, has emerged as a promising solution for Few-Shot Class-Incremental Learning (FSCIL). However, existing methods still struggle to balance base-class discriminability and new-class generalization. Moreover, limited access to original data during incremental tasks often results in ambiguous inter-class decision boundaries. To address these challenges, we propose SMP (Sculpting Margin Penalty), a novel FSCIL method that strategically integrates margin penalties at different stages within the parameter-efficient fine-tuning paradigm. Specifically, we introduce the Margin-aware Intra-task Adapter Merging (MIAM) mechanism for base task learning. MIAM trains two sets of low-rank adapters with distinct classification losses: one with a margin penalty to enhance base-class discriminability, and the other without margin constraints to promote generalization to future new classes. These adapters are then adaptively merged to improve forward compatibility. For incremental tasks, we propose a Margin Penalty-based Classifier Calibration (MPCC) strategy to refine decision boundaries by fine-tuning classifiers on all seen classes' embeddings with a margin penalty. Extensive experiments on CIFAR100, ImageNet-R, and CUB200 demonstrate that SMP achieves state-of-the-art performance in FSCIL while maintaining a better balance between base and new classes.

cs.CV↗

arXiv2025

C-LoRA: Continual Low-Rank Adaptation for Pre-trained Models

Low-Rank Adaptation (LoRA) is an efficient fine-tuning method that has been extensively applied in areas such as natural language processing and computer vision. Existing LoRA fine-tuning approaches excel in static environments but struggle in dynamic learning due to reliance on multiple adapter modules, increasing overhead and complicating inference. We propose Continual Low-Rank Adaptation (C-LoRA), a novel extension of LoRA for continual learning. C-LoRA uses a learnable routing matrix to dynamically manage parameter updates across tasks, ensuring efficient reuse of learned subspaces while enforcing orthogonality to minimize interference and forgetting. Unlike existing approaches that require separate adapters for each task, C-LoRA enables a integrated approach for task adaptation, achieving both scalability and parameter efficiency in sequential learning scenarios. C-LoRA achieves state-of-the-art accuracy and parameter efficiency on benchmarks while providing theoretical insights into its routing matrix's role in retaining and transferring knowledge, establishing a scalable framework for continual learning.

cs.LG↗

arXiv2025

Progressive Local Alignment for Medical Multimodal Pre-training

Local alignment between medical images and text is essential for accurate diagnosis, though it remains challenging due to the absence of natural local pairings and the limitations of rigid region recognition methods. Traditional approaches rely on hard boundaries, which introduce uncertainty, whereas medical imaging demands flexible soft region recognition to handle irregular structures. To overcome these challenges, we propose the Progressive Local Alignment Network (PLAN), which designs a novel contrastive learning-based approach for local alignment to establish meaningful word-pixel relationships and introduces a progressive learning strategy to iteratively refine these relationships, enhancing alignment precision and robustness. By combining these techniques, PLAN effectively improves soft region recognition while suppressing noise interference. Extensive experiments on multiple medical datasets demonstrate that PLAN surpasses state-of-the-art methods in phrase grounding, image-text retrieval, object detection, and zero-shot classification, setting a new benchmark for medical image-text alignment.

cs.CV↗

arXiv2025

ProMedTS: A Self-Supervised, Prompt-Guided Multimodal Approach for Integrating Medical Text and Time Series

Large language models (LLMs) have shown remarkable performance in vision-language tasks, but their application in the medical field remains underexplored, particularly for integrating structured time series data with unstructured clinical notes. In clinical practice, dynamic time series data, such as lab test results, capture critical temporal patterns, while clinical notes provide rich semantic context. Merging these modalities is challenging due to the inherent differences between continuous signals and discrete text. To bridge this gap, we introduce ProMedTS, a novel self-supervised multimodal framework that employs prompt-guided learning to unify these heterogeneous data types. Our approach leverages lightweight anomaly detection to generate anomaly captions that serve as prompts, guiding the encoding of raw time series data into informative prompt embeddings. These prompt embeddings are aligned with textual representations in a shared latent space, preserving fine-grained temporal nuances alongside semantic insights. Furthermore, our framework incorporates tailored self-supervised objectives to enhance both intra- and inter-modal alignment. We evaluate ProMedTS on disease diagnosis tasks using real-world datasets, and the results demonstrate that our method consistently outperforms state-of-the-art approaches.

cs.CL↗

arXiv2024

Knowledge-Augmented Multimodal Clinical Rationale Generation for Disease Diagnosis with Small Language Models

Interpretation is critical for disease diagnosis, but existing models struggle to balance predictive accuracy with human-understandable rationales. While large language models (LLMs) offer strong reasoning abilities, their clinical use is limited by high computational costs and restricted multimodal reasoning ability. Small language models (SLMs) are efficient but lack advanced reasoning for integrating multimodal medical data. In addition, both LLMs and SLMs lack domain knowledge for trustworthy reasoning. Therefore, we propose ClinRaGen, enhancing SLMs by leveraging LLM-derived reasoning ability via rationale distillation and domain knowledge injection for trustworthy multimodal rationale generation. Key innovations include a sequential rationale distillation framework that equips SLMs with LLM-comparable multimodal reasoning abilities, and a knowledge-augmented attention mechanism that jointly unifies multimodal representation from time series and textual data in the same encoding space, enabling it to be naturally interpreted by SLMs while incorporating domain knowledge for reliable rationale generation. Experiments on real-world medical datasets show that ClinRaGen achieves state-of-the-art performance in disease diagnosis and rationale generation, demonstrating the effectiveness of combining LLM-driven reasoning with knowledge augmentation for improved interpretability.

cs.CL↗

arXiv2024

Efficient Non-Exemplar Class-Incremental Learning with Retrospective Feature Synthesis

Despite the outstanding performance in many individual tasks, deep neural networks suffer from catastrophic forgetting when learning from continuous data streams in real-world scenarios. Current Non-Exemplar Class-Incremental Learning (NECIL) methods mitigate forgetting by storing a single prototype per class, which serves to inject previous information when sequentially learning new classes. However, these stored prototypes or their augmented variants often fail to simultaneously capture spatial distribution diversity and precision needed for representing old classes. Moreover, as the model acquires new knowledge, these prototypes gradually become outdated, making them less effective. To overcome these limitations, we propose a more efficient NECIL method that replaces prototypes with synthesized retrospective features for old classes. Specifically, we model each old class's feature space using a multivariate Gaussian distribution and generate deep representations by sampling from high-likelihood regions. Additionally, we introduce a similarity-based feature compensation mechanism that integrates generated old class features with similar new class features to synthesize robust retrospective representations. These retrospective features are then incorporated into our incremental learning framework to preserve the decision boundaries of previous classes while learning new ones. Extensive experiments on CIFAR-100, TinyImageNet, and ImageNet-Subset demonstrate that our method significantly improves the efficiency of non-exemplar class-incremental learning and achieves state-of-the-art performance.

cs.CV↗