Gang Xiao

EAAI Journal 2026 Journal Article

A high-performance defect detection for titanium strip via receptance weighted key value architecture-inspired context modeling and hierarchical differential fusion

• He Zeng
• Y.C. Lin
• Xian-Hua Tan
• Gang Xiao

IJCAI Conference 2025 Conference Paper

Learning to Explain: Towards Human-Aligned Explainability in Deep Reinforcement Learning via Attention Guidance

• Bokai Ji
• Guangxia Li
• Yulong Shen
• Gang Xiao

Recent advances in explainable deep reinforcement learning (DRL) have provided insights into the reasoning behind decisions made by DRL agents. However, existing methods often overlook the subjective nature of explanations and fail to consider human cognitive styles and preferences. Such ignorance tends to reduce the interpretability and relevance of the generated explanations from a human evaluator's perspective. To address this issue, we introduce human cognition into the explaining procedure by integrating DRL with attention guidance in a novel manner. The proposed concept proximal policy optimization (Concept-PPO) learns to generate human-aligned explanations by jointly optimizing the DRL performance and the discrepancy between generated explanations and human annotations. Its key component is a specially designed spatial concept transformer that can enhance explaining efficiency by premasking decision-irrelevant information. Experiments on the ATARI benchmark demonstrate that Concept-PPO achieves better policies than its black-box counterparts, and user studies confirm its superiority in generating human-aligned explanations compared to existing explainable DRL methods.

EAAI Journal 2024 Journal Article

Heterogeneous propagation graph convolution network for a recommendation system based on a knowledge graph

• Jiawei Lu
• Jiapeng Li
• Wenhui Li
• Junfeng Song
• Gang Xiao

AAAI Conference 2023 Conference Paper

Contrastive Predictive Autoencoders for Dynamic Point Cloud Self-Supervised Learning

• Xiaoxiao Sheng
• Zhiqiang Shen
• Gang Xiao

We present a new self-supervised paradigm on point cloud sequence understanding. Inspired by the discriminative and generative self-supervised methods, we design two tasks, namely point cloud sequence based Contrastive Prediction and Reconstruction (CPR), to collaboratively learn more comprehensive spatiotemporal representations. Specifically, dense point cloud segments are first input into an encoder to extract embeddings. All but the last ones are then aggregated by a context-aware autoregressor to make predictions for the last target segment. Towards the goal of modeling multi-granularity structures, local and global contrastive learning are performed between predictions and targets. To further improve the generalization of representations, the predictions are also utilized to reconstruct raw point cloud sequences by a decoder, where point cloud colorization is employed to discriminate against different frames. By combining classic contrast and reconstruction paradigms, it makes the learned representations with both global discrimination and local perception. We conduct experiments on four point cloud sequence benchmarks, and report the results on action recognition and gesture recognition under multiple experimental settings. The performances are comparable with supervised methods and show powerful transferability.

AAAI Conference 2005 Conference Paper

Software Testing by Active Learning for Commercial Games

• Gang Xiao
• Robert C. Holte

As software systems have become larger, exhaustive testing has become increasingly onerous. This has rendered statistical software testing and machine learning techniques increasingly attractive. Drawing from both of these, we present an active learning framework for blackbox software testing. The active learning approach samples input/output pairs from a blackbox and learns a model of the system’s behaviour. This model is then used to select new inputs for sampling. This framework has been developed in the context of commercial video games, complex virtual worlds with highdimensional state spaces, too large for exhaustive testing. Beyond its correctness, developers need to evaluate the gameplay of a game, properties such as difficulty. We use the learned model not only to guide sampling but also to summarize the game’s behaviour for the developer to evaluate. We present results from our semi-automated gameplay analysis by machine learning (SAGA-ML) tool applied to Electronics Arts’ FIFA Soccer game.