arXiv subjects

Chunxiao Li

Publications and source records attributed to Chunxiao Li.

28 records · Page 2Linked to original sources

arXiv2025

Uncertainty-aware Knowledge Tracing

Knowledge Tracing (KT) is crucial in education assessment, which focuses on depicting students' learning states and assessing students' mastery of subjects. With the rise of modern online learning platforms, particularly massive open online courses (MOOCs), an abundance of interaction data has greatly advanced the development of the KT technology. Previous research commonly adopts deterministic representation to capture students' knowledge states, which neglects the uncertainty during student interactions and thus fails to model the true knowledge state in learning process. In light of this, we propose an Uncertainty-Aware Knowledge Tracing model (UKT) which employs stochastic distribution embeddings to represent the uncertainty in student interactions, with a Wasserstein self-attention mechanism designed to capture the transition of state distribution in student learning behaviors. Additionally, we introduce the aleatory uncertainty-aware contrastive learning loss, which strengthens the model's robustness towards different types of uncertainties. Extensive experiments on six real-world datasets demonstrate that UKT not only significantly surpasses existing deep learning-based models in KT prediction, but also shows unique advantages in handling the uncertainty of student interactions.

cs.LG↗

arXiv2024

An Efficient Framework for Enhancing Discriminative Models via Diffusion Techniques

Image classification serves as the cornerstone of computer vision, traditionally achieved through discriminative models based on deep neural networks. Recent advancements have introduced classification methods derived from generative models, which offer the advantage of zero-shot classification. However, these methods suffer from two main drawbacks: high computational overhead and inferior performance compared to discriminative models. Inspired by the coordinated cognitive processes of rapid-slow pathway interactions in the human brain during visual signal recognition, we propose the Diffusion-Based Discriminative Model Enhancement Framework (DBMEF). This framework seamlessly integrates discriminative and generative models in a training-free manner, leveraging discriminative models for initial predictions and endowing deep neural networks with rethinking capabilities via diffusion models. Consequently, DBMEF can effectively enhance the classification accuracy and generalization capability of discriminative models in a plug-and-play manner. We have conducted extensive experiments across 17 prevalent deep model architectures with different training methods, including both CNN-based models such as ResNet and Transformer-based models like ViT, to demonstrate the effectiveness of the proposed DBMEF. Specifically, the framework yields a 1.51\% performance improvement for ResNet-50 on the ImageNet dataset and 3.02\% on the ImageNet-A dataset. In conclusion, our research introduces a novel paradigm for image classification, demonstrating stable improvements across different datasets and neural networks. The code is available at https://github.com/ChunXiaostudy/DBMEF.

cs.CV↗

arXiv2024

Boosting Single Positive Multi-label Classification with Generalized Robust Loss

Multi-label learning (MLL) requires comprehensive multi-semantic annotations that is hard to fully obtain, thus often resulting in missing labels scenarios. In this paper, we investigate Single Positive Multi-label Learning (SPML), where each image is associated with merely one positive label. Existing SPML methods only focus on designing losses using mechanisms such as hard pseudo-labeling and robust losses, mostly leading to unacceptable false negatives. To address this issue, we first propose a generalized loss framework based on expected risk minimization to provide soft pseudo labels, and point out that the former losses can be seamlessly converted into our framework. In particular, we design a novel robust loss based on our framework, which enjoys flexible coordination between false positives and false negatives, and can additionally deal with the imbalance between positive and negative samples. Extensive experiments show that our approach can significantly improve SPML performance and outperform the vast majority of state-of-the-art methods on all the four benchmarks.

cs.LG↗

arXiv2024

A Reinforcement Learning based Reset Policy for CDCL SAT Solvers

Restart policy is an important technique used in modern Conflict-Driven Clause Learning (CDCL) solvers, wherein some parts of the solver state are erased at certain intervals during the run of the solver. In most solvers, variable activities are preserved across restart boundaries, resulting in solvers continuing to search parts of the assignment tree that are not far from the one immediately prior to a restart. To enable the solver to search possibly "distant" parts of the assignment tree, we study the effect of resets, a variant of restarts which not only erases the assignment trail, but also randomizes the activity scores of the variables of the input formula after reset, thus potentially enabling a better global exploration of the search space. In this paper, we model the problem of whether to trigger reset as a multi-armed bandit (MAB) problem, and propose two reinforcement learning (RL) based adaptive reset policies using the Upper Confidence Bound (UCB) and Thompson sampling algorithms. These two algorithms balance the exploration-exploitation tradeoff by adaptively choosing arms (reset vs. no reset) based on their estimated rewards during the solver's run. We implement our reset policies in four baseline SOTA CDCL solvers and compare the baselines against the reset versions on Satcoin benchmarks and SAT Competition instances. Our results show that RL-based reset versions outperform the corresponding baseline solvers on both Satcoin and the SAT competition instances, suggesting that our RL policy helps to dynamically and profitably adapt the reset frequency for any given input instance. We also introduce the concept of a partial reset, where at least a constant number of variable activities are retained across reset boundaries. Building on previous results, we show that there is an exponential separation between O(1) vs. $Ω(n)$-length partial resets.

cs.LO↗

arXiv2023

Boosting Multi-modal Model Performance with Adaptive Gradient Modulation

While the field of multi-modal learning keeps growing fast, the deficiency of the standard joint training paradigm has become clear through recent studies. They attribute the sub-optimal performance of the jointly trained model to the modality competition phenomenon. Existing works attempt to improve the jointly trained model by modulating the training process. Despite their effectiveness, those methods can only apply to late fusion models. More importantly, the mechanism of the modality competition remains unexplored. In this paper, we first propose an adaptive gradient modulation method that can boost the performance of multi-modal models with various fusion strategies. Extensive experiments show that our method surpasses all existing modulation methods. Furthermore, to have a quantitative understanding of the modality competition and the mechanism behind the effectiveness of our modulation method, we introduce a novel metric to measure the competition strength. This metric is built on the mono-modal concept, a function that is designed to represent the competition-less state of a modality. Through systematic investigation, our results confirm the intuition that the modulation encourages the model to rely on the more informative modality. In addition, we find that the jointly trained model typically has a preferred modality on which the competition is weaker than other modalities. However, this preferred modality need not dominate others. Our code will be available at https://github.com/lihong2303/AGM_ICCV2023.

cs.CV↗

arXiv2023

Reconstructing human activities via coupling mobile phone data with location-based social networks

In the era of big data, the ubiquity of location-aware portable devices provides an unprecedented opportunity to understand inhabitants' behavior and their interactions with the built environments. Among the widely used data resources, mobile phone data is the one passively collected and has the largest coverage in the population. However, mobile operators cannot pinpoint one user within meters, leading to the difficulties in activity inference. To that end, we propose a data analysis framework to identify user's activity via coupling the mobile phone data with location-based social networks (LBSN) data. The two datasets are integrated into a Bayesian inference module, considering people's circadian rhythms in both time and space. Specifically, the framework considers the pattern of arrival time to each type of facility and the spatial distribution of facilities. The former can be observed from the LBSN Data and the latter is provided by the points of interest (POIs) dataset. Taking Shanghai as an example, we reconstruct the activity chains of 1,000,000 active mobile phone users and analyze the temporal and spatial characteristics of each activity type. We assess the results with some official surveys and a real-world check-in dataset collected in Shanghai, indicating that the proposed method can capture and analyze human activities effectively. Next, we cluster users' inferred activity chains with a topic model to understand the behavior of different groups of users. This data analysis framework provides an example of reconstructing and understanding the activity of the population at an urban scale with big data fusion.

cs.SI↗

arXiv2023

Limits of CDCL Learning via Merge Resolution

In their seminal work, Atserias et al. and independently Pipatsrisawat and Darwiche in 2009 showed that CDCL solvers can simulate resolution proofs with polynomial overhead. However, previous work does not address the tightness of the simulation, i.e., the question of how large this overhead needs to be. In this paper, we address this question by focusing on an important property of proofs generated by CDCL solvers that employ standard learning schemes, namely that the derivation of a learned clause has at least one inference where a literal appears in both premises (aka, a merge literal). Specifically, we show that proofs of this kind can simulate resolution proofs with at most a linear overhead, but there also exist formulas where such overhead is necessary or, more precisely, that there exist formulas with resolution proofs of linear length that require quadratic CDCL proofs.

cs.CC↗

arXiv2022

Self Supervised Lesion Recognition For Breast Ultrasound Diagnosis

Previous deep learning based Computer Aided Diagnosis (CAD) system treats multiple views of the same lesion as independent images. Since an ultrasound image only describes a partial 2D projection of a 3D lesion, such paradigm ignores the semantic relationship between different views of a lesion, which is inconsistent with the traditional diagnosis where sonographers analyze a lesion from at least two views. In this paper, we propose a multi-task framework that complements Benign/Malignant classification task with lesion recognition (LR) which helps leveraging relationship among multiple views of a single lesion to learn a complete representation of the lesion. To be specific, LR task employs contrastive learning to encourage representation that pulls multiple views of the same lesion and repels those of different lesions. The task therefore facilitates a representation that is not only invariant to the view change of the lesion, but also capturing fine-grained features to distinguish between different lesions. Experiments show that the proposed multi-task framework boosts the performance of Benign/Malignant classification as two sub-tasks complement each other and enhance the learned representation of ultrasound images.

eess.IV↗

arXiv2021

On the Hierarchical Community Structure of Practical Boolean Formulas

Modern CDCL SAT solvers easily solve industrial instances containing tens of millions of variables and clauses, despite the theoretical intractability of the SAT problem. This gap between practice and theory is a central problem in solver research. It is believed that SAT solvers exploit structure inherent in industrial instances, and hence there have been numerous attempts over the last 25 years at characterizing this structure via parameters. These can be classified as rigorous, i.e., they serve as a basis for complexity-theoretic upper bounds (e.g., backdoors), or correlative, i.e., they correlate well with solver run time and are observed in industrial instances (e.g., community structure). Unfortunately, no parameter proposed to date has been shown to be both strongly correlative and rigorous over a large fraction of industrial instances. Given the sheer difficulty of the problem, we aim for an intermediate goal of proposing a set of parameters that is strongly correlative and has good theoretical properties. Specifically, we propose parameters based on a graph partitioning called Hierarchical Community Structure (HCS), which captures the recursive community structure of a graph of a Boolean formula. We show that HCS parameters are strongly correlative with solver run time using an Empirical Hardness Model, and further build a classifier based on HCS parameters that distinguishes between easy industrial and hard random/crafted instances with very high accuracy. We further strengthen our hypotheses via scaling studies. On the theoretical side, we show that counterexamples which plagued community structure do not apply to HCS, and that there is a subset of HCS parameters such that restricting them limits the size of embeddable expanders.

cs.LO↗

arXiv2020

Towards a Complexity-theoretic Understanding of Restarts in SAT solvers

Restarts are a widely-used class of techniques integral to the efficiency of Conflict-Driven Clause Learning (CDCL) Boolean SAT solvers. While the utility of such policies has been well-established empirically, a theoretical explanation of whether restarts are indeed crucial to the power of CDCL solvers is lacking. In this paper, we prove a series of theoretical results that characterize the power of restarts for various models of SAT solvers. More precisely, we make the following contributions. First, we prove an exponential separation between a {\it drunk} randomized CDCL solver model with restarts and the same model without restarts using a family of satisfiable instances. Second, we show that the configuration of CDCL solver with VSIDS branching and restarts (with activities erased after restarts) is exponentially more powerful than the same configuration without restarts for a family of unsatisfiable instances. To the best of our knowledge, these are the first separation results involving restarts in the context of SAT solvers. Third, we show that restarts do not add any proof complexity-theoretic power vis-a-vis a number of models of CDCL and DPLL solvers with non-deterministic static variable and value selection.

cs.CC↗