Guangtao Wang

Guangtao Wang, Jun Li, Zhijian Wu et al.

In recent years, deep convolutional neural networks (CNN) have significantly advanced face detection. In particular, lightweight CNNbased architectures have achieved great success due to their lowcomplexity structure facilitating real-time detection tasks. However, current lightweight CNN-based face detectors trading accuracy for efficiency have inadequate capability in handling insufficient feature representation, faces with unbalanced aspect ratios and occlusion. Consequently, they exhibit deteriorated performance far lagging behind the deep heavy detectors. To achieve efficient face detection without sacrificing accuracy, we design an efficient deep face detector termed EfficientFace in this study, which contains three modules for feature enhancement. To begin with, we design a novel cross-scale feature fusion strategy to facilitate bottom-up information propagation, such that fusing low-level and highlevel features is further strengthened. Besides, this is conducive to estimating the locations of faces and enhancing the descriptive power of face features. Secondly, we introduce a Receptive Field Enhancement module to consider faces with various aspect ratios. Thirdly, we add an Attention Mechanism module for improving the representational capability of occluded faces. We have evaluated EfficientFace on four public benchmarks and experimental results demonstrate the appealing performance of our method. In particular, our model respectively achieves 95.1% (Easy), 94.0% (Medium) and 90.1% (Hard) on validation set of WIDER Face dataset, which is competitive with heavyweight models with only 1/15 computational costs of the state-of-the-art MogFace detector.

Jie Liu, Mengting He, Guangtao Wang et al.

Imbalanced node classification widely exists in real-world networks where graph neural networks (GNNs) are usually highly inclined to majority classes and suffer from severe performance degradation on classifying minority class nodes. Various imbalanced node classification methods have been proposed recently which construct synthetic nodes and edges w.r.t. minority classes to balance the label and topology distribution. However, they are all based on the homophilic assumption that nodes of the same label tend to connect despite the wide existence of heterophilic edges in real-world graphs. Thus, they uniformly aggregate features from both homophilic and heterophilic neighbors and rely on feature similarity to generate synthetic edges, which cannot be applied to imbalanced graphs in high heterophily. To address this problem, we propose a novel GraphSANN for imbalanced node classification on both homophilic and heterophilic graphs. Firstly, we propose a unified feature mixer to generate synthetic nodes with both homophilic and heterophilic interpolation in a unified way. Next, by randomly sampling edges between synthetic nodes and existing nodes as candidate edges, we design an adaptive subgraph extractor to adaptively extract the contextual subgraphs of candidate edges with flexible ranges. Finally, we develop a multi-filter subgraph encoder that constructs different filter channels to discriminatively aggregate neighbor's information along the homophilic and heterophilic edges. Extensive experiments on eight datasets demonstrate the superiority of our model for imbalanced node classification on both homophilic and heterophilic graphs.

Chao Shang, Guangtao Wang, Peng Qi et al.

Question answering over temporal knowledge graphs (KGs) efficiently uses facts contained in a temporal KG, which records entity relations and when they occur in time, to answer natural language questions (e.g., "Who was the president of the US before Obama?"). These questions often involve three time-related challenges that previous work fail to adequately address: 1) questions often do not specify exact timestamps of interest (e.g., "Obama" instead of 2000); 2) subtle lexical differences in time relations (e.g., "before" vs "after"); 3) off-the-shelf temporal KG embeddings that previous work builds on ignore the temporal order of timestamps, which is crucial for answering temporal-order related questions. In this paper, we propose a time-sensitive question answering (TSQA) framework to tackle these problems. TSQA features a timestamp estimation module to infer the unwritten timestamp from the question. We also employ a time-sensitive KG encoder to inject ordering information into the temporal KG embeddings that TSQA is based on. With the help of techniques to reduce the search space for potential answers, TSQA significantly outperforms the previous state of the art on a new benchmark for question answering over temporal KGs, especially achieving a 32% (absolute) error reduction on complex questions that require multiple steps of reasoning over facts in the temporal KG.

Guangtao Wang, Jun Li, Jie Xie et al.

In face detection, low-resolution faces, such as numerous small faces of a human group in a crowded scene, are common in dense face prediction tasks. They usually contain limited visual clues and make small faces less distinguishable from the other small objects, which poses great challenge to accurate face detection. Although deep convolutional neural network has significantly promoted the research on face detection recently, current deep face detectors rarely take into account low-resolution faces and are still vulnerable to the real-world scenarios where massive amount of low-resolution faces exist. Consequently, they usually achieve degraded performance for low-resolution face detection. In order to alleviate this problem, we develop an efficient detector termed EfficientSRFace by introducing a feature-level super-resolution reconstruction network for enhancing the feature representation capability of the model. This module plays an auxiliary role in the training process, and can be removed during the inference without increasing the inference time. Extensive experiments on public benchmarking datasets, such as FDDB and WIDER Face, show that the embedded image super-resolution module can significantly improve the detection accuracy at the cost of a small amount of additional parameters and computational overhead, while helping our model achieve competitive performance compared with the state-of-the-arts methods.

Peng Qi, Guangtao Wang, Jing Huang

Distilling supervision signal from a long sequence to make predictions is a challenging task in machine learning, especially when not all elements in the input sequence contribute equally to the desired output. In this paper, we propose SpanDrop, a simple and effective data augmentation technique that helps models identify the true supervision signal in a long sequence with very few examples. By directly manipulating the input sequence, SpanDrop randomly ablates parts of the sequence at a time and ask the model to perform the same task to emulate counterfactual learning and achieve input attribution. Based on theoretical analysis of its properties, we also propose a variant of SpanDrop based on the beta-Bernoulli distribution, which yields diverse augmented sequences while providing a learning objective that is more consistent with the original dataset. We demonstrate the effectiveness of SpanDrop on a set of carefully designed toy tasks, as well as various natural language processing tasks that require reasoning over long sequences to arrive at the correct answer, and show that it helps models improve performance both when data is scarce and abundant.

Jonathan Li, Nasim Farahini, Evgenii Iuliugin et al.

The proliferation of 100B+ parameter Large Language Models (LLMs) with 100k+ context length support have resulted in increasing demands for on-chip memory to support large KV caches. Techniques such as StreamingLLM and SnapKV demonstrate how to control KV cache size while maintaining model accuracy. Yet, these techniques are not commonly used within industrial deployments using frameworks like vLLM or SGLang. The reason is twofold: on one hand, the static graphs and continuous batching methodology employed by these frameworks make it difficult to admit modifications to the standard multi-head attention algorithm, while on the other hand, the accuracy implications of such techniques on modern instruction-following and reasoning models are not well understood, obfuscating the need for implementing these techniques. In this paper, we explore these accuracy implications on Llama-3.1-8B-Instruct and DeepSeek-R1, and develop SnapStream, a KV cache compression method that can be deployed at scale. We demonstrate the efficacy of SnapStream in a 16-way tensor-parallel deployment of DeepSeek-671B on SambaNova SN40L accelerators running at 128k context length and up to 1832 tokens per second in a real production setting. SnapStream enables $4\times$ improved on-chip memory usage and introduces minimal accuracy degradation on LongBench-v2, AIME24 and LiveCodeBench. To the best of our knowledge, this is the first implementation of sparse KV attention techniques deployed in a production inference system with static graphs and continuous batching.

Shubhangi Upasani, Ravi Shanker Raju, Bo Li et al.

Prompt length is a major bottleneck in agentic large language model (LLM) workloads, where repeated inference steps and multi-call loops incur substantial prefill cost. Recent work on speculative prefill demonstrates that attention-based token importance estimation can enable training-free prompt compression, but this assumes the existence of a draft model that shares the same tokenizer as the target model. In practice, however, agentic pipelines frequently employ models without any smaller in-family draft model. In this work, we study cross-family speculative prefill, where a lightweight draft model from one model family is used to perform prompt compression for a target model from a different family. Using the same speculative prefill mechanism as prior work, we evaluate a range of cross-family draft-target combinations, including Qwen, LLaMA, and DeepSeek models. Across a broad diversity of tasks, we find that attention-based token importance estimation transfers reliably across different model families despite differences in model architectures and tokenizers between draft and target models. Cross-model prompt compression largely retains 90~100% of full-prompt baseline performance and, in some cases, slightly improves accuracy due to denoising effects, while delivering substantial reductions in time to first token (TTFT). These results suggest that speculative prefill depends mainly on task priors and semantic structure, thus serving as a generalizable prompt compression primitive. We discuss the implications of our findings for agentic systems, where repeated long-context inference and heterogeneous model stacks make cross-model prompt compression both necessary and practical.

Guangtao Wang, Shubhangi Upasani, Chen Wu et al.

Efficient long-context inference is critical as large language models (LLMs) adopt context windows of ranging from 128K to 1M tokens. However, the growing key-value (KV) cache and the high computational complexity of attention create significant bottlenecks in memory usage and latency. In this paper, we find that attention in diverse long-context tasks exhibits sparsity, and LLMs implicitly "know" which tokens can be dropped or evicted at the head level after the pre-filling stage. Based on this insight, we propose Self-Attention Guided Eviction~(SAGE-KV), a simple and effective KV eviction cache method for long-context inference. After prefilling, our method performs a one-time top-k selection at both the token and head levels to compress the KV cache, enabling efficient inference with the reduced cache. Evaluations on LongBench and three long-context LLMs (Llama3.1-8B-Instruct-128k, Llama3-8B-Prolong-512k-Instruct, and Qwen2.5-7B-Instruct-128k) show that SAGE-KV maintains accuracy comparable to full attention while significantly improving efficiency. Specifically, SAGE-KV achieves 4x higher memory efficiency with improved accuracy over the static KV cache selection method StreamLLM, and 2x higher memory efficiency with better accuracy than the dynamic KV cache selection method Quest.

Xiaochen Hou, Peng Qi, Guangtao Wang et al.

Recent work on aspect-level sentiment classification has demonstrated the efficacy of incorporating syntactic structures such as dependency trees with graph neural networks(GNN), but these approaches are usually vulnerable to parsing errors. To better leverage syntactic information in the face of unavoidable errors, we propose a simple yet effective graph ensemble technique, GraphMerge, to make use of the predictions from differ-ent parsers. Instead of assigning one set of model parameters to each dependency tree, we first combine the dependency relations from different parses before applying GNNs over the resulting graph. This allows GNN mod-els to be robust to parse errors at no additional computational cost, and helps avoid overparameterization and overfitting from GNN layer stacking by introducing more connectivity into the ensemble graph. Our experiments on the SemEval 2014 Task 4 and ACL 14 Twitter datasets show that our GraphMerge model not only outperforms models with single dependency tree, but also beats other ensemble mod-els without adding model parameters.

Guangtao Wang, Qinbao Song, Xiaoyan Zhu

Recommending appropriate algorithms to a classification problem is one of the most challenging issues in the field of data mining. The existing algorithm recommendation models are generally constructed on only one kind of meta-features by single learners. Considering that i) ensemble learners usually show better performance and ii) different kinds of meta-features characterize the classification problems in different viewpoints independently, and further the models constructed with different sets of meta-features will be complementary with each other and applicable for ensemble. This paper proposes an ensemble learning-based algorithm recommendation method. To evaluate the proposed recommendation method, extensive experiments with 13 well-known candidate classification algorithms and five different kinds of meta-features are conducted on 1090 benchmark classification problems. The results show the effectiveness of the proposed ensemble learning based recommendation method.

Guangtao Wang, Rex Ying, Jing Huang et al.

Self-attention mechanism in graph neural networks (GNNs) led to state-of-the-art performance on many graph representation learning tasks. Currently, at every layer, attention is computed between connected pairs of nodes and depends solely on the representation of the two nodes. However, such attention mechanism does not account for nodes that are not directly connected but provide important network context. Here we propose Multi-hop Attention Graph Neural Network (MAGNA), a principled way to incorporate multi-hop context information into every layer of attention computation. MAGNA diffuses the attention scores across the network, which increases the receptive field for every layer of the GNN. Unlike previous approaches, MAGNA uses a diffusion prior on attention values, to efficiently account for all paths between the pair of disconnected nodes. We demonstrate in theory and experiments that MAGNA captures large-scale structural information in every layer, and has a low-pass effect that eliminates noisy high-frequency information from graph data. Experimental results on node classification as well as the knowledge graph completion benchmarks show that MAGNA achieves state-of-the-art results: MAGNA achieves up to 5.7 percent relative error reduction over the previous state-of-the-art on Cora, Citeseer, and Pubmed. MAGNA also obtains the best performance on a large-scale Open Graph Benchmark dataset. On knowledge graph completion MAGNA advances state-of-the-art on WN18RR and FB15k-237 across four different performance metrics.

Bin Wang, Guangtao Wang, Jing Huang et al.

Commonsense knowledge graph (CKG) is a special type of knowledge graph (KG), where entities are composed of free-form text. However, most existing CKG completion methods focus on the setting where all the entities are presented at training time. Although this setting is standard for conventional KG completion, it has limitations for CKG completion. At test time, entities in CKGs can be unseen because they may have unseen text/names and entities may be disconnected from the training graph, since CKGs are generally very sparse. Here, we propose to study the inductive learning setting for CKG completion where unseen entities may present at test time. We develop a novel learning framework named InductivE. Different from previous approaches, InductiveE ensures the inductive learning capability by directly computing entity embeddings from raw entity attributes/text. InductiveE consists of a free-text encoder, a graph encoder, and a KG completion decoder. Specifically, the free-text encoder first extracts the textual representation of each entity based on the pre-trained language model and word embedding. The graph encoder is a gated relational graph convolutional neural network that learns from a densified graph for more informative entity representation learning. We develop a method that densifies CKGs by adding edges among semantic-related entities and provide more supportive information for unseen entities, leading to better generalization ability of entity embedding for unseen entities. Finally, inductiveE employs Conv-TransE as the CKG completion decoder. Experimental results show that InductiveE significantly outperforms state-of-the-art baselines in both standard and inductive settings on ATOMIC and ConceptNet benchmarks. InductivE performs especially well on inductive scenarios where it achieves above 48% improvement over present methods.

Kevin Huang, Guangtao Wang, Tengyu Ma et al.

Document-level relation extraction is a challenging task which requires reasoning over multiple sentences in order to predict relations in a document. In this paper, we pro-pose a joint training frameworkE2GRE(Entity and Evidence Guided Relation Extraction)for this task. First, we introduce entity-guided sequences as inputs to a pre-trained language model (e.g. BERT, RoBERTa). These entity-guided sequences help a pre-trained language model (LM) to focus on areas of the document related to the entity. Secondly, we guide the fine-tuning of the pre-trained language model by using its internal attention probabilities as additional features for evidence prediction.Our new approach encourages the pre-trained language model to focus on the entities and supporting/evidence sentences. We evaluate our E2GRE approach on DocRED, a recently released large-scale dataset for relation extraction. Our approach is able to achieve state-of-the-art results on the public leaderboard across all metrics, showing that our E2GRE is both effective and synergistic on relation extraction and evidence prediction.

Yun Tang, Jing Huang, Guangtao Wang et al.

Translational distance-based knowledge graph embedding has shown progressive improvements on the link prediction task, from TransE to the latest state-of-the-art RotatE. However, N-1, 1-N and N-N predictions still remain challenging. In this work, we propose a novel translational distance-based approach for knowledge graph link prediction. The proposed method includes two-folds, first we extend the RotatE from 2D complex domain to high dimension space with orthogonal transforms to model relations for better modeling capacity. Second, the graph context is explicitly modeled via two directed context representations. These context representations are used as part of the distance scoring function to measure the plausibility of the triples during training and inference. The proposed approach effectively improves prediction accuracy on the difficult N-1, 1-N and N-N cases for knowledge graph link prediction task. The experimental results show that it achieves better performance on two benchmark data sets compared to the baseline RotatE, especially on data set (FB15k-237) with many high in-degree connection nodes.

Ming Tu, Kevin Huang, Guangtao Wang et al.

Interpretable multi-hop reading comprehension (RC) over multiple documents is a challenging problem because it demands reasoning over multiple information sources and explaining the answer prediction by providing supporting evidences. In this paper, we propose an effective and interpretable Select, Answer and Explain (SAE) system to solve the multi-document RC problem. Our system first filters out answer-unrelated documents and thus reduce the amount of distraction information. This is achieved by a document classifier trained with a novel pairwise learning-to-rank loss. The selected answer-related documents are then input to a model to jointly predict the answer and supporting sentences. The model is optimized with a multi-task learning objective on both token level for answer prediction and sentence level for supporting sentences prediction, together with an attention-based interaction between these two tasks. Evaluated on HotpotQA, a challenging multi-hop RC data set, the proposed SAE system achieves top competitive performance in distractor setting compared to other existing systems on the leaderboard.

Guangtao Wang, Rex Ying, Jing Huang et al.

Graph Attention Networks (GATs) are the state-of-the-art neural architecture for representation learning with graphs. GATs learn attention functions that assign weights to nodes so that different nodes have different influences in the feature aggregation steps. In practice, however, induced attention functions are prone to over-fitting due to the increasing number of parameters and the lack of direct supervision on attention weights. GATs also suffer from over-smoothing at the decision boundary of nodes. Here we propose a framework to address their weaknesses via margin-based constraints on attention during training. We first theoretically demonstrate the over-smoothing behavior of GATs and then develop an approach using constraint on the attention weights according to the class boundary and feature aggregation pattern. Furthermore, to alleviate the over-fitting problem, we propose additional constraints on the graph structure. Extensive experiments and ablation studies on common benchmark datasets demonstrate the effectiveness of our method, which leads to significant improvements over the previous state-of-the-art graph attention methods on all datasets.

Xiaochen Hou, Jing Huang, Guangtao Wang et al.

Aspect-level sentiment classification aims to identify the sentiment polarity towards a specific aspect term in a sentence. Most current approaches mainly consider the semantic information by utilizing attention mechanisms to capture the interactions between the context and the aspect term. In this paper, we propose to employ graph convolutional networks (GCNs) on the dependency tree to learn syntax-aware representations of aspect terms. GCNs often show the best performance with two layers, and deeper GCNs do not bring additional gain due to over-smoothing problem. However, in some cases, important context words cannot be reached within two hops on the dependency tree. Therefore we design a selective attention based GCN block (SA-GCN) to find the most important context words, and directly aggregate these information into the aspect-term representation. We conduct experiments on the SemEval 2014 Task 4 datasets. Our experimental results show that our model outperforms the current state-of-the-art.

Ming Tu, Guangtao Wang, Jing Huang et al.

Multi-hop reading comprehension (RC) across documents poses new challenge over single-document RC because it requires reasoning over multiple documents to reach the final answer. In this paper, we propose a new model to tackle the multi-hop RC problem. We introduce a heterogeneous graph with different types of nodes and edges, which is named as Heterogeneous Document-Entity (HDE) graph. The advantage of HDE graph is that it contains different granularity levels of information including candidates, documents and entities in specific document contexts. Our proposed model can do reasoning over the HDE graph with nodes representation initialized with co-attention and self-attention based context encoders. We employ Graph Neural Networks (GNN) based message passing algorithms to accumulate evidences on the proposed HDE graph. Evaluated on the blind test set of the Qangaroo WikiHop data set, our HDE graph based single model delivers competitive result, and the ensemble model achieves the state-of-the-art performance.

Guangtao Wang, Qinbao Song, Heli Sun et al.

Many feature subset selection (FSS) algorithms have been proposed, but not all of them are appropriate for a given feature selection problem. At the same time, so far there is rarely a good way to choose appropriate FSS algorithms for the problem at hand. Thus, FSS algorithm automatic recommendation is very important and practically useful. In this paper, a meta learning based FSS algorithm automatic recommendation method is presented. The proposed method first identifies the data sets that are most similar to the one at hand by the k-nearest neighbor classification algorithm, and the distances among these data sets are calculated based on the commonly-used data set characteristics. Then, it ranks all the candidate FSS algorithms according to their performance on these similar data sets, and chooses the algorithms with best performance as the appropriate ones. The performance of the candidate FSS algorithms is evaluated by a multi-criteria metric that takes into account not only the classification accuracy over the selected features, but also the runtime of feature selection and the number of selected features. The proposed recommendation method is extensively tested on 115 real world data sets with 22 well-known and frequently-used different FSS algorithms for five representative classifiers. The results show the effectiveness of our proposed FSS algorithm recommendation method.