arXiv subjects

Zhiguang Wang

Publications and source records attributed to Zhiguang Wang.

At least 19 recordsLinked to original sources

arXiv2026

Realizable N:M Sparse Transformer Inference via Search-Kernel Co-Design

Vision Transformers (ViTs) achieve strong accuracy but incur high inference latency. Semi-structured N:M sparsity can reduce arithmetic cost, yet its theoretical savings often fail to translate into proportional end-to-end speedups on modern GPUs. This mismatch arises because deployment latency depends not only on arithmetic reduction but also on execution regularity and hardware scheduling under sparsity. Achieving practical acceleration, therefore, requires coordinated design across sparse execution and sparsity configuration. To this end, we propose a hardware-software co-design framework for N:M sparse ViT inference. On the hardware side, we design MD-SpMM, an N:M sparse CUDA kernel that reorganizes sparse GEMM into micro-dense, Tensor-Core-aligned dataflow and uses inference-aware adaptive parallelism to sustain utilization. On the software side, we perform layer-wise sparsity search under explicit end-to-end latency budgets using a three-stage heuristic search with constraint relaxation to avoid premature convergence and enable deployment-aware sparsity allocation. Experiments on multiple ViT/Swin models and GPU platforms show that the framework achieves over 2.2x latency speedup while maintaining comparable accuracy and delivering superior accuracy under the same latency constraint. The source code is publicly available at https://github.com/liuganhuo/realizable-nm-sparse-transformer.

cs.AR↗

arXiv2026

Time-RA: Towards Time Series Reasoning for Anomaly Diagnosis with LLM Feedback

Time series anomaly detection (TSAD) has traditionally focused on binary classification and often lacks the fine-grained categorization and explanatory reasoning required for transparent decision-making. To address these limitations, we propose Time-series Reasoning for Anomaly (Time-RA), a novel task that reformulates TSAD from a discriminative into a generative, reasoning-intensive paradigm. To facilitate this, we introduce RATs40K, the first real-world large-scale multimodal benchmark with ~40,000 samples across 10 domains, integrating raw time series, textual context, and visual plots with structured reasoning annotations. Extensive benchmarking shows that while supervised fine-tuning and visual representations boost diagnostic accuracy and reasoning consistency, performance varies across complex scenarios. Notably, fine-tuned models demonstrate strong "plug-and-play" transferability, outperforming traditional baselines on unseen real-world datasets. Our work establishes a foundation for interpretable, multimodal time series analysis. All code (https://github.com/yyysjz1997/Time-RA) and the RATs40K dataset (https://huggingface.co/datasets/Time-RA/RATs40K) are fully open-sourced to facilitate future research.

cs.LG↗

arXiv2024

Prove Symbolic Regression is NP-hard by Symbol Graph

Symbolic regression (SR) is the task of discovering a symbolic expression that fits a given data set from the space of mathematical expressions. Despite the abundance of research surrounding the SR problem, there's a scarcity of works that confirm its NP-hard nature. Therefore, this paper introduces the concept of a symbol graph as a comprehensive representation of the entire mathematical expression space, effectively illustrating the NP-hard characteristics of the SR problem. Leveraging the symbol graph, we establish a connection between the SR problem and the task of identifying an optimally fitted degree-constrained Steiner Arborescence (DCSAP). The complexity of DCSAP, which is proven to be NP-hard, directly implies the NP-hard nature of the SR problem.

cs.CC↗

arXiv2023

Adopting Robustness and Optimality in Fitting and Learning

We generalized a modified exponentialized estimator by pushing the robust-optimal (RO) index $λ$ to $-\infty$ for achieving robustness to outliers by optimizing a quasi-Minimin function. The robustness is realized and controlled adaptively by the RO index without any predefined threshold. Optimality is guaranteed by expansion of the convexity region in the Hessian matrix to largely avoid local optima. Detailed quantitative analysis on both robustness and optimality are provided. The results of proposed experiments on fitting tasks for three noisy non-convex functions and the digits recognition task on the MNIST dataset consolidate the conclusions.

cs.LG↗

arXiv2023

Improving Native Ads CTR Prediction by Large Scale Event Embedding and Recurrent Networks

Click through rate (CTR) prediction is very important for Native advertisement but also hard as there is no direct query intent. In this paper we propose a large-scale event embedding scheme to encode the each user browsing event by training a Siamese network with weak supervision on the users' consecutive events. The CTR prediction problem is modeled as a supervised recurrent neural network, which naturally model the user history as a sequence of events. Our proposed recurrent models utilizing pretrained event embedding vectors and an attention layer to model the user history. Our experiments demonstrate that our model significantly outperforms the baseline and some variants.

cs.LG↗

arXiv2022

Taylor Genetic Programming for Symbolic Regression

Genetic programming (GP) is a commonly used approach to solve symbolic regression (SR) problems. Compared with the machine learning or deep learning methods that depend on the pre-defined model and the training dataset for solving SR problems, GP is more focused on finding the solution in a search space. Although GP has good performance on large-scale benchmarks, it randomly transforms individuals to search results without taking advantage of the characteristics of the dataset. So, the search process of GP is usually slow, and the final results could be unstable.To guide GP by these characteristics, we propose a new method for SR, called Taylor genetic programming (TaylorGP) (Code and appendix at https://kgae-cup.github.io/TaylorGP/). TaylorGP leverages a Taylor polynomial to approximate the symbolic equation that fits the dataset. It also utilizes the Taylor polynomial to extract the features of the symbolic equation: low order polynomial discrimination, variable separability, boundary, monotonic, and parity. GP is enhanced by these Taylor polynomial techniques. Experiments are conducted on three kinds of benchmarks: classical SR, machine learning, and physics. The experimental results show that TaylorGP not only has higher accuracy than the nine baseline methods, but also is faster in finding stable results.

cs.NE↗

arXiv2022

Exploring Hidden Semantics in Neural Networks with Symbolic Regression

Many recent studies focus on developing mechanisms to explain the black-box behaviors of neural networks (NNs). However, little work has been done to extract the potential hidden semantics (mathematical representation) of a neural network. A succinct and explicit mathematical representation of a NN model could improve the understanding and interpretation of its behaviors. To address this need, we propose a novel symbolic regression method for neural works (called SRNet) to discover the mathematical expressions of a NN. SRNet creates a Cartesian genetic programming (NNCGP) to represent the hidden semantics of a single layer in a NN. It then leverages a multi-chromosome NNCGP to represent hidden semantics of all layers of the NN. The method uses a (1+$λ$) evolutionary strategy (called MNNCGP-ES) to extract the final mathematical expressions of all layers in the NN. Experiments on 12 symbolic regression benchmarks and 5 classification benchmarks show that SRNet not only can reveal the complex relationships between each layer of a NN but also can extract the mathematical representation of the whole NN. Compared with LIME and MAPLE, SRNet has higher interpolation accuracy and trends to approximate the real model on the practical dataset.

cs.LG↗

arXiv2021

Zero-Shot Dialogue State Tracking via Cross-Task Transfer

Zero-shot transfer learning for dialogue state tracking (DST) enables us to handle a variety of task-oriented dialogue domains without the expense of collecting in-domain data. In this work, we propose to transfer the \textit{cross-task} knowledge from general question answering (QA) corpora for the zero-shot DST task. Specifically, we propose TransferQA, a transferable generative QA model that seamlessly combines extractive QA and multi-choice QA via a text-to-text transformer framework, and tracks both categorical slots and non-categorical slots in DST. In addition, we introduce two effective ways to construct unanswerable questions, namely, negative question sampling and context truncation, which enable our model to handle "none" value slots in the zero-shot DST setting. The extensive experiments show that our approaches substantially improve the existing zero-shot and few-shot results on MultiWoz. Moreover, compared to the fully trained baseline on the Schema-Guided Dialogue dataset, our approach shows better generalization ability in unseen domains.

cs.CL↗

arXiv2021

Leveraging Slot Descriptions for Zero-Shot Cross-Domain Dialogue State Tracking

Zero-shot cross-domain dialogue state tracking (DST) enables us to handle task-oriented dialogue in unseen domains without the expense of collecting in-domain data. In this paper, we propose a slot description enhanced generative approach for zero-shot cross-domain DST. Specifically, our model first encodes dialogue context and slots with a pre-trained self-attentive encoder, and generates slot values in an auto-regressive manner. In addition, we incorporate Slot Type Informed Descriptions that capture the shared information across slots to facilitate cross-domain knowledge transfer. Experimental results on the MultiWOZ dataset show that our proposed method significantly improves existing state-of-the-art results in the zero-shot cross-domain setting.

cs.CL↗

arXiv2021

Adding Chit-Chat to Enhance Task-Oriented Dialogues

Existing dialogue corpora and models are typically designed under two disjoint motives: while task-oriented systems focus on achieving functional goals (e.g., booking hotels), open-domain chatbots aim at making socially engaging conversations. In this work, we propose to integrate both types of systems by Adding Chit-Chat to ENhance Task-ORiented dialogues (ACCENTOR), with the goal of making virtual assistant conversations more engaging and interactive. Specifically, we propose a Human <-> AI collaborative data collection approach for generating diverse chit-chat responses to augment task-oriented dialogues with minimal annotation effort. We then present our new chit-chat-based annotations to 23.8K dialogues from two popular task-oriented datasets (Schema-Guided Dialogue and MultiWOZ 2.1) and demonstrate their advantage over the originals via human evaluation. Lastly, we propose three new models for adding chit-chat to task-oriented dialogues, explicitly trained to predict user goals and to generate contextually relevant chit-chat responses. Automatic and human evaluations show that, compared with the state-of-the-art task-oriented baseline, our models can code-switch between task and chit-chat to be more engaging, interesting, knowledgeable, and humanlike, while maintaining competitive task performance.

cs.CL↗

arXiv2021

Gated Transformer Networks for Multivariate Time Series Classification

Deep learning model (primarily convolutional networks and LSTM) for time series classification has been studied broadly by the community with the wide applications in different domains like healthcare, finance, industrial engineering and IoT. Meanwhile, Transformer Networks recently achieved frontier performance on various natural language processing and computer vision tasks. In this work, we explored a simple extension of the current Transformer Networks with gating, named Gated Transformer Networks (GTN) for the multivariate time series classification problem. With the gating that merges two towers of Transformer which model the channel-wise and step-wise correlations respectively, we show how GTN is naturally and effectively suitable for the multivariate time series classification task. We conduct comprehensive experiments on thirteen dataset with full ablation study. Our results show that GTN is able to achieve competing results with current state-of-the-art deep learning models. We also explored the attention map for the natural interpretability of GTN on time series modeling. Our preliminary results provide a strong baseline for the Transformer Networks on multivariate time series classification task and grounds the foundation for future research.

cs.LG↗

arXiv2020

Continual Learning in Task-Oriented Dialogue Systems

Continual learning in task-oriented dialogue systems can allow us to add new domains and functionalities through time without incurring the high cost of a whole system retraining. In this paper, we propose a continual learning benchmark for task-oriented dialogue systems with 37 domains to be learned continuously in four settings, such as intent recognition, state tracking, natural language generation, and end-to-end. Moreover, we implement and compare multiple existing continual learning baselines, and we propose a simple yet effective architectural method based on residual adapters. Our experiments demonstrate that the proposed architectural method and a simple replay-based strategy perform comparably well but they both achieve inferior performance to the multi-task learning baseline, in where all the data are shown at once, showing that continual learning in task-oriented dialogue systems is a challenging task. Furthermore, we reveal several trade-offs between different continual learning methods in term of parameter usage and memory size, which are important in the design of a task-oriented dialogue system. The proposed benchmark is released together with several baselines to promote more research in this direction.

cs.CL↗

arXiv2020

Information Seeking in the Spirit of Learning: a Dataset for Conversational Curiosity

Open-ended human learning and information-seeking are increasingly mediated by digital assistants. However, such systems often ignore the user's pre-existing knowledge. Assuming a correlation between engagement and user responses such as "liking" messages or asking followup questions, we design a Wizard-of-Oz dialog task that tests the hypothesis that engagement increases when users are presented with facts related to what they know. Through crowd-sourcing of this experiment, we collect and release 14K dialogs (181K utterances) where users and assistants converse about geographic topics like geopolitical entities and locations. This dataset is annotated with pre-existing user knowledge, message-level dialog acts, grounding to Wikipedia, and user reactions to messages. Responses using a user's prior knowledge increase engagement. We incorporate this knowledge into a multi-task model that reproduces human assistant policies and improves over a BERT content model by 13 mean reciprocal rank points.

cs.CL↗

arXiv2019

Diabetic Retinopathy Detection via Deep Convolutional Networks for Discriminative Localization and Visual Explanation

We proposed a deep learning method for interpretable diabetic retinopathy (DR) detection. The visual-interpretable feature of the proposed method is achieved by adding the regression activation map (RAM) after the global averaging pooling layer of the convolutional networks (CNN). With RAM, the proposed model can localize the discriminative regions of an retina image to show the specific region of interest in terms of its severity level. We believe this advantage of the proposed deep learning model is highly desired for DR detection because in practice, users are not only interested with high prediction performance, but also keen to understand the insights of DR detection and why the adopted learning model works. In the experiments conducted on a large scale of retina image dataset, we show that the proposed CNN model can achieve high performance on DR detection compared with the state-of-the-art while achieving the merits of providing the RAM to highlight the salient regions of the input image.

cs.CV↗

arXiv2019

Voltage Control of Perpendicular Exchange Bias in Multiferroic Heterostructures

Perpendicular exchange bias (EB), which combines the perpendicular magnetic anisotropy and the ferromagnetic (FM) - antiferromagnetic (AFM) exchange coupling, is extremely important in the high-density AFM spintronics. However, the effective modulation of EB remains challenging, since the alternant spins at the AFM/FM interface are strongly pinned by the AFM layer. Voltage tuning of EB through the magnetoelectric coupling provides a potential way to achieve a rapid magnetization switching in an energy-efficient manner. Nevertheless, the interfacial strain mediation of perpendicular EB induced by E-field remains unexplored. In this work, we obtain perpendicular EB nanostructure by room-temperature fabrication process, and demonstrate the voltage tunable perpendicular EB in Pt/IrMn/(Co/Pt)2/Ta/(011) Pb(Mg1/3Nb2/3)O3-PbTiO3 multiferroic heterostructure. To enhance the voltage control effect on perpendicular EB, we further investigate both strain-mediated magnetoelectric coupling and ionic liquid gating method in the thinned EB system with the structure of Pt/IrMn/Co/Pt/Ta. As a result, the voltage induced lattice distortion effectively transmits to the AFM/FM interface, while the charge accumulation in gating method generates a relatively large hysteresis loop offset that has not been observed before at room temperature. The voltage manipulation of perpendicular EB at room temperature provides new possibilities towards novel AFM devices and memories with great energy-efficiency and ultra-high density.

cond-mat.mtrl-sci↗

arXiv2018

Encoding Temporal Markov Dynamics in Graph for Visualizing and Mining Time Series

Time series and signals are attracting more attention across statistics, machine learning and pattern recognition as it appears widely in the industry especially in sensor and IoT related research and applications, but few advances has been achieved in effective time series visual analytics and interaction due to its temporal dimensionality and complex dynamics. Inspired by recent effort on using network metrics to characterize time series for classification, we present an approach to visualize time series as complex networks based on the first order Markov process in its temporal ordering. In contrast to the classical bar charts, line plots and other statistics based graph, our approach delivers more intuitive visualization that better preserves both the temporal dependency and frequency structures. It provides a natural inverse operation to map the graph back to raw signals, making it possible to use graph statistics to characterize time series for better visual exploration and statistical analysis. Our experimental results suggest the effectiveness on various tasks such as pattern discovery and classification on both synthetic and the real time series and sensor data.

cs.LG↗

arXiv2017

Automated Cloud Provisioning on AWS using Deep Reinforcement Learning

As the use of cloud computing continues to rise, controlling cost becomes increasingly important. Yet there is evidence that 30\% - 45\% of cloud spend is wasted. Existing tools for cloud provisioning typically rely on highly trained human experts to specify what to monitor, thresholds for triggering action, and actions. In this paper we explore the use of reinforcement learning (RL) to acquire policies to balance performance and spend, allowing humans to specify what they want as opposed to how to do it, minimizing the need for cloud expertise. Empirical results with tabular, deep, and dueling double deep Q-learning with the CloudSim simulator show the utility of RL and the relative merits of the approaches. We also demonstrate effective policy transfer learning from an extremely simple simulator to CloudSim, with the next step being transfer from CloudSim to an Amazon Web Services physical environment.

cs.DC↗

arXiv2016

Time Series Classification from Scratch with Deep Neural Networks: A Strong Baseline

We propose a simple but strong baseline for time series classification from scratch with deep neural networks. Our proposed baseline models are pure end-to-end without any heavy preprocessing on the raw data or feature crafting. The proposed Fully Convolutional Network (FCN) achieves premium performance to other state-of-the-art approaches and our exploration of the very deep neural networks with the ResNet structure is also competitive. The global average pooling in our convolutional model enables the exploitation of the Class Activation Map (CAM) to find out the contributing region in the raw data for the specific labels. Our models provides a simple choice for the real world application and a good starting point for the future research. An overall analysis is provided to discuss the generalization capability of our models, learned features, network structures and the classification semantics.

cs.LG↗