Haoran Li

Haoran Li, Yang Weng, Hanghang Tong

Learning the underlying equation from data is a fundamental problem in many disciplines. Recent advances rely on Neural Networks (NNs) but do not provide theoretical guarantees in obtaining the exact equations owing to the non-convexity of NNs. In this paper, we propose Convex Neural Symbolic Learning (CoNSoLe) to seek convexity under mild conditions. The main idea is to decompose the recovering process into two steps and convexify each step. In the first step of searching for right symbols, we convexify the deep Q-learning. The key is to maintain double convexity for both the negative Q-function and the negative reward function in each iteration, leading to provable convexity of the negative optimal Q function to learn the true symbol connections. Conditioned on the exact searching result, we construct a Locally Convex equation Learner (LoCaL) neural network to convexify the estimation of symbol coefficients. With such a design, we quantify a large region with strict convexity in the loss surface of LoCaL for commonly used physical functions. Finally, we demonstrate the superior performance of the CoNSoLe framework over the state-of-the-art on a diverse set of datasets.

Haoran Li, XiaoLu Li, Yihang Lin et al.

Video prediction is a complex time-series forecasting task with great potential in many use cases. However, traditional methods prioritize accuracy and overlook slow prediction speeds due to complex model structures, redundant information, and excessive GPU memory consumption. These methods often predict frames sequentially, making acceleration difficult and limiting their applicability in real-time scenarios like danger prediction and warning.Therefore, we propose a transformer-based keypoint prediction neural network (TKN). TKN extracts dynamic content from video frames in an unsupervised manner, reducing redundant feature computation. And, TKN uses an acceleration matrix to reduce the computational cost of attention and employs a parallel computing structure for prediction acceleration. To the best of our knowledge, TKN is the first real-time video prediction solution that achieves a prediction rate of 1,176 fps, significantly reducing computation costs while maintaining other performance. Qualitative and quantitative experiments on multiple datasets have demonstrated the superiority of our method, suggesting that TKN has great application potential.

Haoran Li, Jiayu Lv, Congying Han et al.

Achieving optimality and adversarial robustness in deep reinforcement learning has long been regarded as conflicting goals. Nonetheless, recent theoretical insights presented in CAR suggest a potential alignment, raising the important question of how to realize this in practice. This paper first identifies a key gap between theory and practice by comparing standard policy optimization (SPO) and adversarially robust policy optimization (ARPO). Although they share theoretical consistency, a fundamental tension between robustness and optimality arises in practical policy gradient methods. SPO tends toward convergence to vulnerable first-order stationary policies (FOSPs) with strong natural performance, whereas ARPO typically favors more robust FOSPs at the expense of reduced returns. Furthermore, we attribute this tradeoff to the reshaping effect of the strongest adversary in ARPO, which significantly complicates the global landscape by inducing deceptive sticky FOSPs. This improves robustness but makes navigation more challenging. To alleviate this, we develop the BARPO, a bilevel framework unifying SPO and ARPO by modulating adversary strength, thereby facilitating navigability while preserving global optima. Extensive empirical results demonstrate that BARPO consistently outperforms vanilla ARPO, providing a practical approach to reconcile theoretical and empirical performance.

Yonggan Fu, Wuyang Chen, Haotao Wang et al.

The compression of Generative Adversarial Networks (GANs) has lately drawn attention, due to the increasing demand for deploying GANs into mobile devices for numerous applications such as image translation, enhancement and editing. However, compared to the substantial efforts to compressing other deep models, the research on compressing GANs (usually the generators) remains at its infancy stage. Existing GAN compression algorithms are limited to handling specific GAN architectures and losses. Inspired by the recent success of AutoML in deep compression, we introduce AutoML to GAN compression and develop an AutoGAN-Distiller (AGD) framework. Starting with a specifically designed efficient search space, AGD performs an end-to-end discovery for new efficient generators, given the target computational resource constraints. The search is guided by the original GAN model via knowledge distillation, therefore fulfilling the compression. AGD is fully automatic, standalone (i.e., needing no trained discriminators), and generically applicable to various GAN models. We evaluate AGD in two representative GAN tasks: image translation and super resolution. Without bells and whistles, AGD yields remarkably lightweight yet more competitive compressed models, that largely outperform existing alternatives. Our codes and pretrained models are available at https://github.com/TAMU-VITA/AGD.

Haoran Li, Haolin Shi, Wenli Zhang et al.

Text-to-3D scene generation holds immense potential for the gaming, film, and architecture sectors. Despite significant progress, existing methods struggle with maintaining high quality, consistency, and editing flexibility. In this paper, we propose DreamScene, a 3D Gaussian-based novel text-to-3D scene generation framework, to tackle the aforementioned three challenges mainly via two strategies. First, DreamScene employs Formation Pattern Sampling (FPS), a multi-timestep sampling strategy guided by the formation patterns of 3D objects, to form fast, semantically rich, and high-quality representations. FPS uses 3D Gaussian filtering for optimization stability, and leverages reconstruction techniques to generate plausible textures. Second, DreamScene employs a progressive three-stage camera sampling strategy, specifically designed for both indoor and outdoor settings, to effectively ensure object-environment integration and scene-wide 3D consistency. Last, DreamScene enhances scene editing flexibility by integrating objects and environments, enabling targeted adjustments. Extensive experiments validate DreamScene's superiority over current state-of-the-art techniques, heralding its wide-ranging potential for diverse applications. Code and demos will be released at https://dreamscene-project.github.io .

Guangzheng Hu, Yuanheng Zhu, Haoran Li et al.

Many real-world applications involve some agents that fall into two teams, with payoffs that are equal within the same team but of opposite sign across the opponent team. The so-called two-team zero-sum Markov games (2t0sMGs) can be resolved with reinforcement learning in recent years. However, existing methods are thus inefficient in light of insufficient consideration of intra-team credit assignment, data utilization and computational intractability. In this paper, we propose the individual-global-minimax (IGMM) principle to ensure the coherence between two-team minimax behaviors and the individual greedy behaviors through Q functions in 2t0sMGs. Based on it, we present a novel multi-agent reinforcement learning framework, Factorized Multi-Agent MiniMax Q-Learning (FM3Q), which can factorize the joint minimax Q function into individual ones and iteratively solve for the IGMM-satisfied minimax Q functions for 2t0sMGs. Moreover, an online learning algorithm with neural networks is proposed to implement FM3Q and obtain the deterministic and decentralized minimax policies for two-team players. A theoretical analysis is provided to prove the convergence of FM3Q. Empirically, we use three environments to evaluate the learning efficiency and final performance of FM3Q and show its superiority on 2t0sMGs.

Anqi Li, Congying Han, Tiande Guo et al.

Existing methods provide varying algorithms for different types of Boolean satisfiability problems (SAT), lacking a general solution framework. Accordingly, this study proposes a unified framework DCSAT based on integer programming and reinforcement learning (RL) algorithm to solve different types of SAT problems such as MaxSAT, Weighted MaxSAT, PMS, WPMS. Specifically, we first construct a consolidated integer programming representation for four types of SAT problems by adjusting objective function coefficients. Secondly, we construct an appropriate reinforcement learning models based on the 0-1 integer programming for SAT problems. Based on the binary tree search structure, we apply the Monte Carlo tree search (MCTS) method on SAT problems. Finally, we prove that this method can find all optimal Boolean assignments based on Wiener-khinchin law of large Numbers. We experimentally verify that this paradigm can prune the unnecessary search space to find the optimal Boolean assignments for the problem. Furthermore, the proposed method can provide diverse labels for supervised learning methods for SAT problems.

Haoran Li, Zicheng Zhang, Wang Luo et al.

Establishing robust policies is essential to counter attacks or disturbances affecting deep reinforcement learning (DRL) agents. Recent studies explore state-adversarial robustness and suggest the potential lack of an optimal robust policy (ORP), posing challenges in setting strict robustness constraints. This work further investigates ORP: At first, we introduce a consistency assumption of policy (CAP) stating that optimal actions in the Markov decision process remain consistent with minor perturbations, supported by empirical and theoretical evidence. Building upon CAP, we crucially prove the existence of a deterministic and stationary ORP that aligns with the Bellman optimal policy. Furthermore, we illustrate the necessity of $L^{\infty}$-norm when minimizing Bellman error to attain ORP. This finding clarifies the vulnerability of prior DRL algorithms that target the Bellman optimal policy with $L^{1}$-norm and motivates us to train a Consistent Adversarial Robust Deep Q-Network (CAR-DQN) by minimizing a surrogate of Bellman Infinity-error. The top-tier performance of CAR-DQN across various benchmarks validates its practical effectiveness and reinforces the soundness of our theoretical analysis.

Wei Xu, Haoran Li, Baoyuan Ou et al.

Cross-domain Click-Through Rate prediction aims to tackle the data sparsity and the cold start problems in online advertising systems by transferring knowledge from source domains to a target domain. Most existing methods rely on overlapping users to facilitate this transfer, often focusing on joint training or pre-training with fine-tuning approach to connect the source and target domains. However, in real-world industrial settings, joint training struggles to learn optimal representations with different distributions, and pre-training with fine-tuning is not well-suited for continuously integrating new data. To address these issues, we propose GIST, a cross-domain lifelong sequence model that decouples the training processes of the source and target domains. Unlike previous methods that search lifelong sequences in the source domains using only content or behavior signals or their simple combinations, we innovatively introduce a Content-Behavior Joint Training Module (CBJT), which aligns content-behavior distributions and combines them with guided information to facilitate a more stable representation. Furthermore, we develop an Asymmetric Similarity Integration strategy (ASI) to augment knowledge transfer through similarity computation. Extensive experiments demonstrate the effectiveness of GIST, surpassing SOTA methods on offline evaluations and an online A/B test. Deployed on the Xiaohongshu (RedNote) platform, GIST effectively enhances online ads system performance at scale, serving hundreds of millions of daily active users.

Wenzhao Liu, Haoran Li, Congying Han et al.

Achieving generalization in neural approaches across different scales and distributions remains a significant challenge for the Traveling Salesman Problem~(TSP). A key obstacle is that neural networks often fail to learn robust principles for identifying universal patterns and deriving optimal solutions from diverse instances. In this paper, we first uncover Purity Law (PuLa), a fundamental structural principle for optimal TSP solutions, defining that edge prevalence grows exponentially with the sparsity of surrounding vertices. Statistically validated across diverse instances, PuLa reveals a consistent bias toward local sparsity in global optima. Building on this insight, we propose Purity Policy Optimization~(PUPO), a novel training paradigm that explicitly aligns characteristics of neural solutions with PuLa during the solution construction process to enhance generalization. Extensive experiments demonstrate that PUPO can be seamlessly integrated with popular neural solvers, significantly enhancing their generalization performance without incurring additional computational overhead during inference.

Haoran Li, Zicheng Zhang, Wang Luo et al.

Ensuring the robustness of deep reinforcement learning (DRL) agents against adversarial attacks is critical for their trustworthy deployment. Recent research highlights the challenges of achieving state-adversarial robustness and suggests that an optimal robust policy (ORP) does not always exist, complicating the enforcement of strict robustness constraints. In this paper, we further explore the concept of ORP. We first introduce the Intrinsic State-adversarial Markov Decision Process (ISA-MDP), a novel formulation where adversaries cannot fundamentally alter the intrinsic nature of state observations. ISA-MDP, supported by empirical and theoretical evidence, universally characterizes decision-making under state-adversarial paradigms. We rigorously prove that within ISA-MDP, a deterministic and stationary ORP exists, aligning with the Bellman optimal policy. Our findings theoretically reveal that improving DRL robustness does not necessarily compromise performance in natural environments. Furthermore, we demonstrate the necessity of infinity measurement error (IME) in both $Q$-function and probability spaces to achieve ORP, unveiling vulnerabilities of previous DRL algorithms that rely on $1$-measurement errors. Motivated by these insights, we develop the Consistent Adversarial Robust Reinforcement Learning (CAR-RL) framework, which optimizes surrogates of IME. We apply CAR-RL to both value-based and policy-based DRL algorithms, achieving superior performance and validating our theoretical analysis.

Chi Zhou, Wang Luo, Haoran Li et al.

Offline reinforcement learning agents face significant deployment challenges due to the synthetic-to-real distribution mismatch. While most prior research has focused on improving the fidelity of synthetic sampling and incorporating off-policy mechanisms, the directly integrated paradigm often fails to ensure consistent policy behavior in biased models and underlying environmental dynamics, which inherently arise from discrepancies between behavior and learning policies. In this paper, we first shift the focus from model reliability to policy discrepancies while optimizing for expected returns, and then self-consistently incorporate synthetic data, deriving a novel actor-critic paradigm, Dual Alignment Maximin Optimization (DAMO). It is a unified framework to ensure both model-environment policy consistency and synthetic and offline data compatibility. The inner minimization performs dual conservative value estimation, aligning policies and trajectories to avoid out-of-distribution states and actions, while the outer maximization ensures that policy improvements remain consistent with inner value estimates. Empirical evaluations demonstrate that DAMO effectively ensures model and policy alignments, achieving competitive performance across diverse benchmark tasks.

Fei Sun, Minghai Qin, Tianyun Zhang et al.

Deep neural networks (DNNs) have been proven to be effective in solving many real-life problems, but its high computation cost prohibits those models from being deployed to edge devices. Pruning, as a method to introduce zeros to model weights, has shown to be an effective method to provide good trade-offs between model accuracy and computation efficiency, and is a widely-used method to generate compressed models. However, the granularity of pruning makes important trade-offs. At the same sparsity level, a coarse-grained structured sparse pattern is more efficient on conventional hardware but results in worse accuracy, while a fine-grained unstructured sparse pattern can achieve better accuracy but is inefficient on existing hardware. On the other hand, some modern processors are equipped with fast on-chip scratchpad memories and gather/scatter engines that perform indirect load and store operations on such memories. In this work, we propose a set of novel sparse patterns, named gather-scatter (GS) patterns, to utilize the scratchpad memories and gather/scatter engines to speed up neural network inferences. Correspondingly, we present a compact sparse format. The proposed set of sparse patterns, along with a novel pruning methodology, address the load imbalance issue and result in models with quality close to unstructured sparse models and computation efficiency close to structured sparse models. Our experiments show that GS patterns consistently make better trade-offs between accuracy and computation efficiency compared to conventional structured sparse patterns. GS patterns can reduce the runtime of the DNN components by two to three times at the same accuracy levels. This is confirmed on three different deep learning tasks and popular models, namely, GNMT for machine translation, ResNet50 for image recognition, and Japser for acoustic speech recognition.

Jiaqi Li, Haoran Li, Yaran Chen et al.

Currently, an increasing number of model pruning methods are proposed to resolve the contradictions between the computer powers required by the deep learning models and the resource-constrained devices. However, most of the traditional rule-based network pruning methods can not reach a sufficient compression ratio with low accuracy loss and are time-consuming as well as laborious. In this paper, we propose Automatic Block-wise and Channel-wise Network Pruning (ABCP) to jointly search the block-wise and channel-wise pruning action with deep reinforcement learning. A joint sample algorithm is proposed to simultaneously generate the pruning choice of each residual block and the channel pruning ratio of each convolutional layer from the discrete and continuous search space respectively. The best pruning action taking both the accuracy and the complexity of the model into account is obtained finally. Compared with the traditional rule-based pruning method, this pipeline saves human labor and achieves a higher compression ratio with lower accuracy loss. Tested on the mobile robot detection dataset, the pruned YOLOv3 model saves 99.5% FLOPs, reduces 99.5% parameters, and achieves 37.3 times speed up with only 2.8% mAP loss. The results of the transfer task on the sim2real detection dataset also show that our pruned model has much better robustness performance.