Shangzhe Li

Junran Wu, Xueyuan Chen, Ke Xu et al.

Following the success of convolution on non-Euclidean space, the corresponding pooling approaches have also been validated on various tasks regarding graphs. However, because of the fixed compression quota and stepwise pooling design, these hierarchical pooling methods still suffer from local structure damage and suboptimal problem. In this work, inspired by structural entropy, we propose a hierarchical pooling approach, SEP, to tackle the two issues. Specifically, without assigning the layer-specific compression quota, a global optimization algorithm is designed to generate the cluster assignment matrices for pooling at once. Then, we present an illustration of the local structure damage from previous methods in the reconstruction of ring and grid synthetic graphs. In addition to SEP, we further design two classification models, SEP-G and SEP-N for graph classification and node classification, respectively. The results show that SEP outperforms state-of-the-art graph pooling methods on graph classification benchmarks and obtains superior performance on node classifications.

Junran Wu, Shangzhe Li, Jianhao Li et al.

In deep neural networks, better results can often be obtained by increasing the complexity of previously developed basic models. However, it is unclear whether there is a way to boost performance by decreasing the complexity of such models. Intuitively, given a problem, a simpler data structure comes with a simpler algorithm. Here, we investigate the feasibility of improving graph classification performance while simplifying the learning process. Inspired by structural entropy on graphs, we transform the data sample from graphs to coding trees, which is a simpler but essential structure for graph data. Furthermore, we propose a novel message passing scheme, termed hierarchical reporting, in which features are transferred from leaf nodes to root nodes by following the hierarchical structure of coding trees. We then present a tree kernel and a convolutional network to implement our scheme for graph classification. With the designed message passing scheme, the tree kernel and convolutional network have a lower runtime complexity of $O(n)$ than Weisfeiler-Lehman subtree kernel and other graph neural networks of at least $O(hm)$. We empirically validate our methods with several graph classification benchmarks and demonstrate that they achieve better performance and lower computational consumption than competing approaches.

Xinming Gao, Shangzhe Li, Yujin Cai et al.

Offline reinforcement learning (RL) enables policy learning from fixed datasets without further environment interaction, making it particularly valuable in high-risk or costly domains. Extreme $Q$-Learning (XQL) is a recent offline RL method that models Bellman errors using the Extreme Value Theorem, yielding strong empirical performance. However, XQL and its stabilized variant MXQL suffer from notable limitations: both require extensive hyperparameter tuning specific to each dataset and domain, and also exhibit instability during training. To address these issues, we proposed a principled method to estimate the temperature coefficient $β$ via quantile regression under mild assumptions. To further improve training stability, we introduce a value regularization technique with mild generalization, inspired by recent advances in constrained value learning. Experimental results demonstrate that the proposed algorithm achieves competitive or superior performance across a range of benchmark tasks, including D4RL and NeoRL2, while maintaining stable training dynamics and using a consistent set of hyperparameters across all datasets and domains.

Yongtao Qu, Shangzhe Li, Weitong Zhang

We consider the offline imitation learning from observations (LfO) where the expert demonstrations are scarce and the available offline suboptimal data are far from the expert behavior. Many existing distribution-matching approaches struggle in this regime because they impose strict support constraints and rely on brittle one-step models, making it hard to extract useful signal from imperfect data. To tackle this challenge, we propose TGE, a trajectory-level generative embedding for offline LfO that constructs a dense, smooth surrogate reward by estimating expert state density in the latent space of a temporal diffusion model trained on offline trajectory data. By leveraging the smooth geometry of the learned diffusion embedding, TGE captures long-horizon temporal dynamics and effectively bridges the gap between disjoint supports, ensuring a robust learning signal even when offline data is distributionally distinct from the expert. Empirically, the proposed approach consistently matches or outperforms prior offline LfO methods across a range of D4RL locomotion and manipulation benchmarks.

Shangzhe Li, Weitong Zhang

We study value adaptation in offline-to-online reinforcement learning under general function approximation. Starting from an imperfect offline pretrained $Q$-function, the learner aims to adapt it to the target environment using only a limited amount of online interaction. We first characterize the difficulty of this setting by establishing a minimax lower bound, showing that even when the pretrained $Q$-function is close to optimal $Q^\star$, online adaptation can be no more efficient than pure online RL on certain hard instances. On the positive side, under a novel structural condition on the offline-pretrained value functions, we propose O2O-LSVI, an adaptation algorithm with problem-dependent sample complexity that provably improves over pure online RL. Finally, we complement our theory with neural-network experiments that demonstrate the practical effectiveness of the proposed method.

He Zhu, Junran Wu, Ruomei Liu et al.

Existing self-supervised methods in natural language processing (NLP), especially hierarchical text classification (HTC), mainly focus on self-supervised contrastive learning, extremely relying on human-designed augmentation rules to generate contrastive samples, which can potentially corrupt or distort the original information. In this paper, we tend to investigate the feasibility of a contrastive learning scheme in which the semantic and syntactic information inherent in the input sample is adequately reserved in the contrastive samples and fused during the learning process. Specifically, we propose an information lossless contrastive learning strategy for HTC, namely \textbf{H}ierarchy-aware \textbf{I}nformation \textbf{L}ossless contrastive \textbf{L}earning (HILL), which consists of a text encoder representing the input document, and a structure encoder directly generating the positive sample. The structure encoder takes the document embedding as input, extracts the essential syntactic information inherent in the label hierarchy with the principle of structural entropy minimization, and injects the syntactic information into the text representation via hierarchical representation learning. Experiments on three common datasets are conducted to verify the superiority of HILL.

Shangzhe Li, Zhiao Huang, Hao Su

Imitation learning (IL) enables agents to acquire skills directly from expert demonstrations, providing a compelling alternative to reinforcement learning. However, prior online IL approaches struggle with complex tasks characterized by high-dimensional inputs and complex dynamics. In this work, we propose a novel approach to online imitation learning that leverages reward-free world models. Our method learns environmental dynamics entirely in latent spaces without reconstruction, enabling efficient and accurate modeling. We adopt the inverse soft-Q learning objective, reformulating the optimization process in the Q-policy space to mitigate the instability associated with traditional optimization in the reward-policy space. By employing a learned latent dynamics model and planning for control, our approach consistently achieves stable, expert-level performance in tasks with high-dimensional observation or action spaces and intricate dynamics. We evaluate our method on a diverse set of benchmarks, including DMControl, MyoSuite, and ManiSkill2, demonstrating superior empirical performance compared to existing approaches.

Xueyuan Chen, Shangzhe Li, Ruomei Liu et al.

Trapped by the label scarcity in molecular property prediction and drug design, graph contrastive learning (GCL) came forward. Leading contrastive learning works show two kinds of view generators, that is, random or learnable data corruption and domain knowledge incorporation. While effective, the two ways also lead to molecular semantics altering and limited generalization capability, respectively. To this end, we relate the \textbf{L}in\textbf{E} graph with \textbf{MO}lecular graph co\textbf{N}trastive learning and propose a novel method termed \textit{LEMON}. Specifically, by contrasting the given graph with the corresponding line graph, the graph encoder can freely encode the molecular semantics without omission. Furthermore, we present a new patch with edge attribute fusion and two local contrastive losses enhance information transmission and tackle hard negative samples. Compared with state-of-the-art (SOTA) methods for view generation, superior performance on molecular property prediction suggests the effectiveness of our proposed framework.

Junran Wu, Xueyuan Chen, Shangzhe Li

Graph contrastive learning has achieved great success in pre-training graph neural networks without ground-truth labels. Leading graph contrastive learning follows the classical scheme of contrastive learning, forcing model to identify the essential information from augmented views. However, general augmented views are produced via random corruption or learning, which inevitably leads to semantics alteration. Although domain knowledge guided augmentations alleviate this issue, the generated views are domain specific and undermine the generalization. In this work, motivated by the firm representation ability of sparse model from pruning, we reformulate the problem of graph contrastive learning via contrasting different model versions rather than augmented views. We first theoretically reveal the superiority of model pruning in contrast to data augmentations. In practice, we take original graph as input and dynamically generate a perturbed graph encoder to contrast with the original encoder by pruning its transformation weights. Furthermore, considering the integrity of node embedding in our method, we are capable of developing a local contrastive loss to tackle the hard negative samples that disturb the model training. We extensively validate our method on various benchmarks regarding graph classification via unsupervised and transfer learning. Compared to the state-of-the-art (SOTA) works, better performance can always be obtained by the proposed method.

Shangzhe Li, Xinhua Zhang

Batch offline data have been shown considerably beneficial for reinforcement learning. Their benefit is further amplified by upsampling with generative models. In this paper, we consider a novel opportunity where interaction with environment is feasible, but only restricted to observations, i.e., \textit{no reward} feedback is available. This setting is broadly applicable, as simulators or even real cyber-physical systems are often accessible, while in contrast reward is often difficult or expensive to obtain. As a result, the learner must make good sense of the offline data to synthesize an efficient scheme of querying the transition of state. Our method first leverages online interactions to generate high-return trajectories via conditional diffusion models. They are then blended with the original offline trajectories through a stitching algorithm, and the resulting augmented data can be applied generically to downstream reinforcement learners. Superior empirical performance is demonstrated over state-of-the-art data augmentation methods that are extended to utilize state-only interactions.

Shangzhe Li, Zhiao Huang, Hao Su

Imitation Learning (IL) has achieved remarkable success across various domains, including robotics, autonomous driving, and healthcare, by enabling agents to learn complex behaviors from expert demonstrations. However, existing IL methods often face instability challenges, particularly when relying on adversarial reward or value formulations in world model frameworks. In this work, we propose a novel approach to online imitation learning that addresses these limitations through a reward model based on random network distillation (RND) for density estimation. Our reward model is built on the joint estimation of expert and behavioral distributions within the latent space of the world model. We evaluate our method across diverse benchmarks, including DMControl, Meta-World, and ManiSkill2, showcasing its ability to deliver stable performance and achieve expert-level results in both locomotion and manipulation tasks. Our approach demonstrates improved stability over adversarial methods while maintaining expert-level performance.

Shangzhe Li, Xuchao Zhang, Chetan Bansal et al.

Self-play post-training methods has emerged as an effective approach for finetuning large language models and turn the weak language model into strong language model without preference data. However, the theoretical foundations for self-play finetuning remain underexplored. In this work, we tackle this by connecting self-play finetuning with adversarial imitation learning by formulating finetuning procedure as a min-max game between the model and a regularized implicit reward player parameterized by the model itself. This perspective unifies self-play imitation and general preference alignment within a common framework. Under this formulation, we present a game-theoretic analysis showing that the self-play finetuning will converge to it's equilibrium. Guided by this theoretical formulation, we propose a new self-play imitation finetuning algorithm based on the $χ^2$-divergence variational objective with bounded rewards and improved stability. Experiments on various of language model finetuning tasks demonstrate consistent improvements over existing self-play methods and validate our theoretical insights.

Shangzhe Li, Dongruo Zhou, Weitong Zhang

We study online adversarial imitation learning (AIL), where an agent learns from offline expert demonstrations and interacts with the environment online without access to rewards. Despite strong empirical results, the benefits of online interaction and the impact of stochasticity remain poorly understood. We address these gaps by introducing a model-based AIL algorithm (MB-AIL) and establish its horizon-free, second-order sample-complexity guarantees under general function approximations for both expert data and reward-free interactions. These second-order bounds provide an instance-dependent result that can scale with the variance of returns under the relevant policies and therefore tighten as the system approaches determinism. Together with second-order, information-theoretic lower bounds on a newly constructed hard-instance family, we show that MB-AIL attains minimax-optimal sample complexity for online interaction (up to logarithmic factors) with limited expert demonstrations and matches the lower bound for expert demonstrations in terms of the dependence on horizon $H$, precision $ε$ and the policy variance $σ^2$. Experiments further validate our theoretical findings and demonstrate that a practical implementation of MB-AIL matches or surpasses the sample efficiency of existing methods.

Zishun Yu, Shangzhe Li, Xinhua Zhang

Large language models have led to significant progress across many NLP tasks, although their massive sizes often incur substantial computational costs. Distillation has become a common practice to compress these large and highly capable models into smaller, more efficient ones. Many existing language model distillation methods can be viewed as behavior cloning from the perspective of imitation learning or inverse reinforcement learning. This viewpoint has inspired subsequent studies that leverage (inverse) reinforcement learning techniques, including variations of behavior cloning and temporal difference learning methods. Rather than proposing yet another specific temporal difference method, we introduce a general framework for temporal difference-based distillation by exploiting the distributional sparsity of the teacher model. Specifically, it is often observed that language models assign most probability mass to a small subset of tokens. Motivated by this observation, we design a temporal difference learning framework that operates on a reduced action space (a subset of vocabulary), and demonstrate how practical algorithms can be derived and the resulting performance improvements.

Junran Wu, Xueyuan Chen, Bowen Shi et al.

In contrastive learning, the choice of ``view'' controls the information that the representation captures and influences the performance of the model. However, leading graph contrastive learning methods generally produce views via random corruption or learning, which could lead to the loss of essential information and alteration of semantic information. An anchor view that maintains the essential information of input graphs for contrastive learning has been hardly investigated. In this paper, based on the theory of graph information bottleneck, we deduce the definition of this anchor view; put differently, \textit{the anchor view with essential information of input graph is supposed to have the minimal structural uncertainty}. Furthermore, guided by structural entropy, we implement the anchor view, termed \textbf{SEGA}, for graph contrastive learning. We extensively validate the proposed anchor view on various benchmarks regarding graph classification under unsupervised, semi-supervised, and transfer learning and achieve significant performance boosts compared to the state-of-the-art methods.

Junran Wu, Ke Xu, Xueyuan Chen et al.

Great research efforts have been devoted to exploiting deep neural networks in stock prediction. While long-range dependencies and chaotic property are still two major issues that lower the performance of state-of-the-art deep learning models in forecasting future price trends. In this study, we propose a novel framework to address both issues. Specifically, in terms of transforming time series into complex networks, we convert market price series into graphs. Then, structural information, referring to associations among temporal points and the node weights, is extracted from the mapped graphs to resolve the problems regarding long-range dependencies and the chaotic property. We take graph embeddings to represent the associations among temporal points as the prediction model inputs. Node weights are used as a priori knowledge to enhance the learning of temporal attention. The effectiveness of our proposed framework is validated using real-world stock data, and our approach obtains the best performance among several state-of-the-art benchmarks. Moreover, in the conducted trading simulations, our framework further obtains the highest cumulative profits. Our results supplement the existing applications of complex network methods in the financial realm and provide insightful implications for investment applications regarding decision support in financial markets.