arXiv subjects

Wei Cheng

Publications and source records attributed to Wei Cheng.

At least 163 records · Page 9Linked to original sources

arXiv2024

Mastering Long-Tail Complexity on Graphs: Characterization, Learning, and Generalization

In the context of long-tail classification on graphs, the vast majority of existing work primarily revolves around the development of model debiasing strategies, intending to mitigate class imbalances and enhance the overall performance. Despite the notable success, there is very limited literature that provides a theoretical tool for characterizing the behaviors of long-tail classes in graphs and gaining insight into generalization performance in real-world scenarios. To bridge this gap, we propose a generalization bound for long-tail classification on graphs by formulating the problem in the fashion of multi-task learning, i.e., each task corresponds to the prediction of one particular class. Our theoretical results show that the generalization performance of long-tail classification is dominated by the overall loss range and the task complexity. Building upon the theoretical findings, we propose a novel generic framework HierTail for long-tail classification on graphs. In particular, we start with a hierarchical task grouping module that allows us to assign related tasks into hypertasks and thus control the complexity of the task space; then, we further design a balanced contrastive learning module to adaptively balance the gradients of both head and tail classes to control the loss range across all tasks in a unified fashion. Extensive experiments demonstrate the effectiveness of HierTail in characterizing long-tail classes on real graphs, which achieves up to 12.9% improvement over the leading baseline method in accuracy.

cs.LG↗

arXiv2024

Dataflow-Guided Retrieval Augmentation for Repository-Level Code Completion

Recent years have witnessed the deployment of code language models (LMs) in various code intelligence tasks such as code completion. Yet, it is challenging for pre-trained LMs to generate correct completions in private repositories. Previous studies retrieve cross-file context based on import relations or text similarity, which is insufficiently relevant to completion targets. In this paper, we propose a dataflow-guided retrieval augmentation approach, called DraCo, for repository-level code completion. DraCo parses a private repository into code entities and establishes their relations through an extended dataflow analysis, forming a repo-specific context graph. Whenever triggering code completion, DraCo precisely retrieves relevant background knowledge from the repo-specific context graph and generates well-formed prompts to query code LMs. Furthermore, we construct a large Python dataset, ReccEval, with more diverse completion targets. Our experiments demonstrate the superior accuracy and applicable efficiency of DraCo, improving code exact match by 3.43% and identifier F1-score by 3.27% on average compared to the state-of-the-art approach.

cs.SE↗

arXiv2024

Pruning as a Domain-specific LLM Extractor

Large Language Models (LLMs) have exhibited remarkable proficiency across a wide array of NLP tasks. However, the escalation in model size also engenders substantial deployment costs. While few efforts have explored model pruning techniques to reduce the size of LLMs, they mainly center on general or task-specific weights. This leads to suboptimal performance due to lacking specificity on the target domain or generality on different tasks when applied to domain-specific challenges. This work introduces an innovative unstructured dual-pruning methodology, D-Pruner, for domain-specific compression on LLM. It extracts a compressed, domain-specific, and task-agnostic LLM by identifying LLM weights that are pivotal for general capabilities, like linguistic capability and multi-task solving, and domain-specific knowledge. More specifically, we first assess general weight importance by quantifying the error incurred upon their removal with the help of an open-domain calibration dataset. Then, we utilize this general weight importance to refine the training loss, so that it preserves generality when fitting into a specific domain. Moreover, by efficiently approximating weight importance with the refined training loss on a domain-specific calibration dataset, we obtain a pruned model emphasizing generality and specificity. Our comprehensive experiments across various tasks in healthcare and legal domains show the effectiveness of D-Pruner in domain-specific compression. Our code is available at https://github.com/psunlpgroup/D-Pruner.

cs.CL↗

arXiv2024

Domain Specialization as the Key to Make Large Language Models Disruptive: A Comprehensive Survey

Large language models (LLMs) have significantly advanced the field of natural language processing (NLP), providing a highly useful, task-agnostic foundation for a wide range of applications. However, directly applying LLMs to solve sophisticated problems in specific domains meets many hurdles, caused by the heterogeneity of domain data, the sophistication of domain knowledge, the uniqueness of domain objectives, and the diversity of the constraints (e.g., various social norms, cultural conformity, religious beliefs, and ethical standards in the domain applications). Domain specification techniques are key to make large language models disruptive in many applications. Specifically, to solve these hurdles, there has been a notable increase in research and practices conducted in recent years on the domain specialization of LLMs. This emerging field of study, with its substantial potential for impact, necessitates a comprehensive and systematic review to better summarize and guide ongoing work in this area. In this article, we present a comprehensive survey on domain specification techniques for large language models, an emerging direction critical for large language model applications. First, we propose a systematic taxonomy that categorizes the LLM domain-specialization techniques based on the accessibility to LLMs and summarizes the framework for all the subcategories as well as their relations and differences to each other. Second, we present an extensive taxonomy of critical application domains that can benefit dramatically from specialized LLMs, discussing their practical significance and open challenges. Last, we offer our insights into the current research status and future trends in this area.

cs.CL↗

arXiv2024

Uncertainty Quantification for In-Context Learning of Large Language Models

In-context learning has emerged as a groundbreaking ability of Large Language Models (LLMs) and revolutionized various fields by providing a few task-relevant demonstrations in the prompt. However, trustworthy issues with LLM's response, such as hallucination, have also been actively discussed. Existing works have been devoted to quantifying the uncertainty in LLM's response, but they often overlook the complex nature of LLMs and the uniqueness of in-context learning. In this work, we delve into the predictive uncertainty of LLMs associated with in-context learning, highlighting that such uncertainties may stem from both the provided demonstrations (aleatoric uncertainty) and ambiguities tied to the model's configurations (epistemic uncertainty). We propose a novel formulation and corresponding estimation method to quantify both types of uncertainties. The proposed method offers an unsupervised way to understand the prediction of in-context learning in a plug-and-play fashion. Extensive experiments are conducted to demonstrate the effectiveness of the decomposition. The code and data are available at: https://github.com/lingchen0331/UQ_ICL.

cs.CL↗

arXiv2024

Towards Robust Fidelity for Evaluating Explainability of Graph Neural Networks

Graph Neural Networks (GNNs) are neural models that leverage the dependency structure in graphical data via message passing among the graph nodes. GNNs have emerged as pivotal architectures in analyzing graph-structured data, and their expansive application in sensitive domains requires a comprehensive understanding of their decision-making processes -- necessitating a framework for GNN explainability. An explanation function for GNNs takes a pre-trained GNN along with a graph as input, to produce a `sufficient statistic' subgraph with respect to the graph label. A main challenge in studying GNN explainability is to provide fidelity measures that evaluate the performance of these explanation functions. This paper studies this foundational challenge, spotlighting the inherent limitations of prevailing fidelity metrics, including $Fid_+$, $Fid_-$, and $Fid_Δ$. Specifically, a formal, information-theoretic definition of explainability is introduced and it is shown that existing metrics often fail to align with this definition across various statistical scenarios. The reason is due to potential distribution shifts when subgraphs are removed in computing these fidelity measures. Subsequently, a robust class of fidelity measures are introduced, and it is shown analytically that they are resilient to distribution shift issues and are applicable in a wide range of scenarios. Extensive empirical analysis on both synthetic and real datasets are provided to illustrate that the proposed metrics are more coherent with gold standard metrics. The source code is available at https://trustai4s-lab.github.io/fidelity.

cs.LG↗

arXiv2024

Constraints on real scalar inflation from preheating of LATTICEEASY

In this paper, we undertake a detailed study of real scalar inflation by employing LATTICEEASY simulations to investigate preheating phenomena. Generally, the scalar inflation potential with non-minimal coupling can be approximated by a quartic potential. We observe that the evolutionary behavior of this potential remains unaffected by the coupling coefficient. Furthermore, the theoretical predictions for the scalar spectral index ($n_s$) and tensor-to-scalar power ratio (r) are also independent of this coefficient. Consequently, the coefficients of this model are not constrained by Planck observations. Fortunately, the properties of preheating after inflation provide a viable approach to examine these coefficients. Through LATTICEEASY simulations, we trace the evolution of particle number density, scale factor, and energy density during the preheating process. Subsequently, we derive the parameters, such as the energy ratio ($γ$) and the e-folding number of preheating ($N_{pre}$), which facilitate further predictions of $n_s$ and $r$. We have successfully validated real scalar inflation model using preheating of LATTICEEASY simulations based on the analytical relationship between preheating and inflation models.

astro-ph.CO↗

arXiv2024

Parametric Augmentation for Time Series Contrastive Learning

Modern techniques like contrastive learning have been effectively used in many areas, including computer vision, natural language processing, and graph-structured data. Creating positive examples that assist the model in learning robust and discriminative representations is a crucial stage in contrastive learning approaches. Usually, preset human intuition directs the selection of relevant data augmentations. Due to patterns that are easily recognized by humans, this rule of thumb works well in the vision and language domains. However, it is impractical to visually inspect the temporal structures in time series. The diversity of time series augmentations at both the dataset and instance levels makes it difficult to choose meaningful augmentations on the fly. In this study, we address this gap by analyzing time series data augmentation using information theory and summarizing the most commonly adopted augmentations in a unified format. We then propose a contrastive learning framework with parametric augmentation, AutoTCL, which can be adaptively employed to support time series representation learning. The proposed approach is encoder-agnostic, allowing it to be seamlessly integrated with different backbone encoders. Experiments on univariate forecasting tasks demonstrate the highly competitive results of our method, with an average 6.5\% reduction in MSE and 4.7\% in MAE over the leading baselines. In classification tasks, AutoTCL achieves a $1.2\%$ increase in average accuracy.

cs.LG↗

arXiv2024

PAC Learnability under Explanation-Preserving Graph Perturbations

Graphical models capture relations between entities in a wide range of applications including social networks, biology, and natural language processing, among others. Graph neural networks (GNN) are neural models that operate over graphs, enabling the model to leverage the complex relationships and dependencies in graph-structured data. A graph explanation is a subgraph which is an `almost sufficient' statistic of the input graph with respect to its classification label. Consequently, the classification label is invariant, with high probability, to perturbations of graph edges not belonging to its explanation subgraph. This work considers two methods for leveraging such perturbation invariances in the design and training of GNNs. First, explanation-assisted learning rules are considered. It is shown that the sample complexity of explanation-assisted learning can be arbitrarily smaller than explanation-agnostic learning. Next, explanation-assisted data augmentation is considered, where the training set is enlarged by artificially producing new training samples via perturbation of the non-explanation edges in the original training set. It is shown that such data augmentation methods may improve performance if the augmented data is in-distribution, however, it may also lead to worse sample complexity compared to explanation-agnostic learning rules if the augmented data is out-of-distribution. Extensive empirical evaluations are provided to verify the theoretical analysis.

cs.LG↗

arXiv2024

Optimal transport in the frame of abstract Lax-Oleinik operator revisited

This is our first paper on the extension of our recent work on the Lax-Oleinik commutators and its applications to the intrinsic approach of propagation of singularities of the viscosity solutions of Hamilton-Jacobi equations. We reformulate Kantorovich-Rubinstein duality theorem in the theory of optimal transport in terms of abstract Lax-Oleinik operators, and analyze the relevant optimal transport problem in the case the cost function $c(x,y)=h(t_1,t_2,x,y)$ is the fundamental solution of Hamilton-Jacobi equation. For further applications to the problem of cut locus and propagation of singularities in optimal transport, we introduce corresponding random Lax-Oleinik operators. We also study the problem of singularities for $c$-concave functions and its dynamical implication when $c$ is the fundamental solution with $t_2-t_1\ll1$ and $t_2-t_1<\infty$, and $c$ is the Peierls' barrier respectively.

math.AP↗

arXiv2024

Experimental test of the Crooks fluctuation theorem in a single nuclear spin

We experimentally test the Crooks fluctuation theorem in a quantum spin system. Our results show that the Crooks fluctuation theorem is valid for different speeds of the nonequilibrium processes and under various effective temperatures. Work is not an observable in quantum systems, which makes tests of quantum thermodynamic theorems challenging. In this work, we developed high-fidelity single-shot readouts of a single nuclear spin in diamond and implemented the two-point work measurement protocol, enabling a direct experimental test of the Crooks fluctuation theorem. Our results provide a quantum insight into fluctuations and the methods we developed can be utilized to study other quantum thermodynamic theorems.

quant-ph↗

arXiv2024

Interpretable Imitation Learning with Dynamic Causal Relations

Imitation learning, which learns agent policy by mimicking expert demonstration, has shown promising results in many applications such as medical treatment regimes and self-driving vehicles. However, it remains a difficult task to interpret control policies learned by the agent. Difficulties mainly come from two aspects: 1) agents in imitation learning are usually implemented as deep neural networks, which are black-box models and lack interpretability; 2) the latent causal mechanism behind agents' decisions may vary along the trajectory, rather than staying static throughout time steps. To increase transparency and offer better interpretability of the neural agent, we propose to expose its captured knowledge in the form of a directed acyclic causal graph, with nodes being action and state variables and edges denoting the causal relations behind predictions. Furthermore, we design this causal discovery process to be state-dependent, enabling it to model the dynamics in latent causal graphs. Concretely, we conduct causal discovery from the perspective of Granger causality and propose a self-explainable imitation learning framework, {\method}. The proposed framework is composed of three parts: a dynamic causal discovery module, a causality encoding module, and a prediction module, and is trained in an end-to-end manner. After the model is learned, we can obtain causal relations among states and action variables behind its decisions, exposing policies learned by it. Experimental results on both synthetic and real-world datasets demonstrate the effectiveness of the proposed {\method} in learning the dynamic causal graphs for understanding the decision-making of imitation learning meanwhile maintaining high prediction accuracy.

cs.LG↗

arXiv2024

Exploring the Impact of Dissipation Coefficient in Warm Higgs Inflation

In this study, we conducted a detailed analysis of the core parameter of Warm Higgs Inflation (WHI) $-$ the dissipation coefficient ($Q$). As a crucial parameter in the warm inflation process, $Q$ exerts profound influences on the entire evolutionary process. By meticulously deriving the relationships between various quantities and $Q$, we successfully circumvented the common preconceptions regarding strong and weak dissipation, laying the foundation for a more accurate exploration of their interconnections. Taking into account the constraints imposed by Cosmic Microwave Background, we observed that the dissipation coefficient $Q$ remains at extremely low levels throughout the entire warm inflation process, i.e., $Q \ll 1$. This observation indicates that WHI falls under the category of weakly dissipative warm inflation. Despite being weakly dissipative, $Q$ still plays a crucial role in the evolution of temperature, energy, and other quantities, highlighting its significance and non-negligibility. We delved deeper into the impact of the primordial power spectrum on the dissipation coefficient $Q$ during the warm inflation process, discovering that the dependency is not significant. Consequently, this naturally leads to the unobtrusive dependence of the gravitational wave power spectrum on $Q$. Finally, we found that gravitational waves generated by WHI hold the potential for verification in future observational experiments, especially through the SKA100 experiment. These findings provide a theoretical support for a more profound understanding of the early evolution of the universe.

astro-ph.CO↗

arXiv2024

Towards Robust Pruning: An Adaptive Knowledge-Retention Pruning Strategy for Language Models

The pruning objective has recently extended beyond accuracy and sparsity to robustness in language models. Despite this, existing methods struggle to enhance robustness against adversarial attacks when continually increasing model sparsity and require a retraining process. As humans step into the era of large language models, these issues become increasingly prominent. This paper proposes that the robustness of language models is proportional to the extent of pre-trained knowledge they encompass. Accordingly, we introduce a post-training pruning strategy designed to faithfully replicate the embedding space and feature space of dense language models, aiming to conserve more pre-trained knowledge during the pruning process. In this setup, each layer's reconstruction error not only originates from itself but also includes cumulative error from preceding layers, followed by an adaptive rectification. Compared to other state-of-art baselines, our approach demonstrates a superior balance between accuracy, sparsity, robustness, and pruning cost with BERT on datasets SST2, IMDB, and AGNews, marking a significant stride towards robust pruning in language models.

cs.CL↗

arXiv2023

Topological and control theoretic properties of Hamilton-Jacobi equations via Lax-Oleinik commutators

In the context of weak KAM theory, we discuss the commutators $\{T^-_t\circ T^+_t\}_{t\geqslant0}$ and $\{T^+_t\circ T^-_t\}_{t\geqslant0}$ of Lax-Oleinik operators. We characterize the relation $T^-_t\circ T^+_t=Id$ for both small time and arbitrary time $t$. We show this relation characterizes controllability for evolutionary Hamilton-Jacobi equation. Based on our previous work on the cut locus of viscosity solution, we refine our analysis of the cut time function $τ$ in terms of commutators $T^+_t\circ T^-_t-T^+_t\circ T^-_t$ and clarify the structure of the super/sub-level set of the cut time function $τ$.

math.AP↗

arXiv2023

Open-ended Commonsense Reasoning with Unrestricted Answer Scope

Open-ended Commonsense Reasoning is defined as solving a commonsense question without providing 1) a short list of answer candidates and 2) a pre-defined answer scope. Conventional ways of formulating the commonsense question into a question-answering form or utilizing external knowledge to learn retrieval-based methods are less applicable in the open-ended setting due to an inherent challenge. Without pre-defining an answer scope or a few candidates, open-ended commonsense reasoning entails predicting answers by searching over an extremely large searching space. Moreover, most questions require implicit multi-hop reasoning, which presents even more challenges to our problem. In this work, we leverage pre-trained language models to iteratively retrieve reasoning paths on the external knowledge base, which does not require task-specific supervision. The reasoning paths can help to identify the most precise answer to the commonsense question. We conduct experiments on two commonsense benchmark datasets. Compared to other approaches, our proposed method achieves better performance both quantitatively and qualitatively.

cs.CL↗

arXiv2023

Experimental test of the Jarzynski equality in a single spin-1 system using high-fidelity single-shot readouts

The Jarzynski equality (JE), which connects the equilibrium free energy with non-equilibrium work statistics, plays a crucial role in quantum thermodynamics. Although practical quantum systems are usually multi-level systems, most tests of the JE were executed in two-level systems. A rigorous test of the JE by directly measuring the work distribution of a physical process in a high-dimensional quantum system remains elusive. Here, we report an experimental test of the JE in a single spin-1 system. We realized nondemolition projective measurement of this three-level system via cascading high-fidelity single-shot readouts and directly measured the work distribution utilizing the two-point measurement protocol. The validity of the JE was verified from the non-adiabatic to adiabatic zone and under different effective temperatures. Our work puts the JE on a solid experimental foundation and makes the NV center system a mature toolbox to perform advanced experiments of stochastic quantum thermodynamics.

quant-ph↗

arXiv2023

DyExplainer: Explainable Dynamic Graph Neural Networks

Graph Neural Networks (GNNs) resurge as a trending research subject owing to their impressive ability to capture representations from graph-structured data. However, the black-box nature of GNNs presents a significant challenge in terms of comprehending and trusting these models, thereby limiting their practical applications in mission-critical scenarios. Although there has been substantial progress in the field of explaining GNNs in recent years, the majority of these studies are centered on static graphs, leaving the explanation of dynamic GNNs largely unexplored. Dynamic GNNs, with their ever-evolving graph structures, pose a unique challenge and require additional efforts to effectively capture temporal dependencies and structural relationships. To address this challenge, we present DyExplainer, a novel approach to explaining dynamic GNNs on the fly. DyExplainer trains a dynamic GNN backbone to extract representations of the graph at each snapshot, while simultaneously exploring structural relationships and temporal dependencies through a sparse attention technique. To preserve the desired properties of the explanation, such as structural consistency and temporal continuity, we augment our approach with contrastive learning techniques to provide priori-guided regularization. To model longer-term temporal dependencies, we develop a buffer-based live-updating scheme for training. The results of our extensive experiments on various datasets demonstrate the superiority of DyExplainer, not only providing faithful explainability of the model predictions but also significantly improving the model prediction accuracy, as evidenced in the link prediction task.

cs.LG↗