Shoubo Hu

Xinwei Shen, Shengyu Zhu, Jiji Zhang et al.

In a nonparametric setting, the causal structure is often identifiable only up to Markov equivalence, and for the purpose of causal inference, it is useful to learn a graphical representation of the Markov equivalence class (MEC). In this paper, we revisit the Greedy Equivalence Search (GES) algorithm, which is widely cited as a score-based algorithm for learning the MEC of the underlying causal structure. We observe that in order to make the GES algorithm consistent in a nonparametric setting, it is not necessary to design a scoring metric that evaluates graphs. Instead, it suffices to plug in a consistent estimator of a measure of conditional dependence to guide the search. We therefore present a reframing of the GES algorithm, which is more flexible than the standard score-based version and readily lends itself to the nonparametric setting with a general measure of conditional dependence. In addition, we propose a neural conditional dependence (NCD) measure, which utilizes the expressive power of deep neural networks to characterize conditional independence in a nonparametric manner. We establish the optimality of the reframed GES algorithm under standard assumptions and the consistency of using our NCD estimator to decide conditional independence. Together these results justify the proposed approach. Experimental results demonstrate the effectiveness of our method in causal discovery, as well as the advantages of using our NCD measure over kernel-based measures.

Wenlong Lyu, Shoubo Hu, Jie Chuai et al.

Bayesian optimization (BO) is widely adopted in black-box optimization problems and it relies on a surrogate model to approximate the black-box response function. With the increasing number of black-box optimization tasks solved and even more to solve, the ability to learn from multiple prior tasks to jointly pre-train a surrogate model is long-awaited to further boost optimization efficiency. In this paper, we propose a simple approach to pre-train a surrogate, which is a Gaussian process (GP) with a kernel defined on deep features learned from a Transformer-based encoder, using datasets from prior tasks with possibly heterogeneous input spaces. In addition, we provide a simple yet effective mix-up initialization strategy for input tokens corresponding to unseen input variables and therefore accelerate new tasks' convergence. Experiments on both synthetic and real benchmark problems demonstrate the effectiveness of our proposed pre-training and transfer BO strategy over existing methods.

Zhentao Tang, Yuqi Cui, Shixiong Kai et al.

Expert-level scientific reasoning remains challenging for large language models, particularly on benchmarks such as Humanity's Last Exam (HLE), where rigid tool pipelines, brittle multi-agent coordination, and inefficient test-time scaling often limit performance. We introduce ReThinker, a confidence-aware agentic framework that orchestrates retrieval, tool use, and multi-agent reasoning through a stage-wise Solver-Critic-Selector architecture. Rather than following a fixed pipeline, ReThinker dynamically allocates computation based on model confidence, enabling adaptive tool invocation, guided multi-dimensional reflection, and robust confidence-weighted selection. To support scalable training without human annotation, we further propose a reverse data synthesis pipeline and an adaptive trajectory recycling strategy that transform successful reasoning traces into high-quality supervision. Experiments on HLE, GAIA, and XBench demonstrate that ReThinker consistently outperforms state-of-the-art foundation models with tools and existing deep research systems, achieving state-of-the-art results on expert-level reasoning tasks.

Qingyu Song, Peiyu Liao, Wenqian Zhao et al.

The increasing deployment of Large Language Models (LLMs) on edge devices, driven by model advancements and hardware improvements, offers significant privacy benefits. However, these on-device LLMs inherently face performance limitations due to reduced model capacity and necessary compression techniques. To address this, we introduce a systematic methodology -- encompassing model capability, development efficiency, and system resources -- for evaluating on-device LLMs. Our comprehensive evaluation, encompassing models from 0.5B to 14B parameters and seven post-training quantization (PTQ) methods on commodity laptops, yields several critical insights: 1) System-level metrics exhibit near-linear scaling with effective bits-per-weight (BPW). 2) A practical threshold exists around $\sim$3.5 effective BPW, larger models subjected to low-bit quantization consistently outperform smaller models utilizing higher bit-precision. 3) Quantization with low BPW incurs marginal accuracy loss but significant memory savings. 4) Determined by low-level implementation specifics power consumption on CPU, where computation-intensive operations spend more power than memory-intensive ones. These findings offer crucial insights and practical guidelines for the efficient deployment and optimized configuration of LLMs on resource-constrained edge devices. Our codebase is available at https://github.com/simmonssong/LLMOnDevice.

Zixiao Wang, Jieya Zhou, Su Zheng et al.

Recent decades have witnessed remarkable advancements in artificial intelligence (AI), including large language models (LLMs), image and video generative models, and embodied AI systems. These advancements have led to an explosive increase in the demand for computational power, challenging the limits of Moore's Law. Optical lithography, a critical technology in semiconductor manufacturing, faces significant challenges due to its high costs. To address this, various lithography simulators have been developed. However, many of these simulators are limited by their inadequate photoresist modeling capabilities. This paper presents TorchResist, an open-source, differentiable photoresist simulator.TorchResist employs an analytical approach to model the photoresist process, functioning as a white-box system with at most twenty interpretable parameters. Leveraging modern differentiable programming techniques and parallel computing on GPUs, TorchResist enables seamless co-optimization with other tools across multiple related tasks. Our experimental results demonstrate that TorchResist achieves superior accuracy and efficiency compared to existing solutions. The source code is publicly available.

Ziru Liu, Cheng Gong, Xinyu Fu et al.

Reinforcement Learning with Verifiable Rewards (RLVR) has recently emerged as a powerful paradigm for facilitating the self-improvement of large language models (LLMs), particularly in the domain of complex reasoning tasks. However, prevailing on-policy RL methods often contend with significant training instability and inefficiency. This is primarily due to a capacity-difficulty mismatch, where the complexity of training data frequently outpaces the model's current capabilities, leading to critically sparse reward signals and stalled learning progress. This challenge is particularly acute for smaller, more resource-efficient LLMs. To overcome this, we introduce the Guided Hybrid Policy Optimization (GHPO), a novel difficulty-aware reinforcement learning framework. GHPO dynamically calibrates task difficulty by employing adaptive prompt refinement to provide targeted guidance. This unique approach adaptively balances direct imitation learning for problems currently beyond the model's reach with exploration-based reinforcement learning for more manageable tasks, effectively creating a smooth and optimized learning curriculum. Extensive experiments demonstrate that GHPO achieves an average performance gain of approximately 5% across six challenging mathematics benchmarks, consistently outperforming strong on-policy reinforcement learning and curriculum learning baselines. Further analysis confirms that our framework significantly enhances both training stability and final reasoning performance, thus offering a scalable and efficient solution for developing powerful and robust reasoning models.

Haoyuan Wu, Xueyi Chen, Rui Ming et al.

Large language models (LLMs) demonstrate significant reasoning capabilities, particularly through long chain-of-thought (CoT) processes, which can be elicited by reinforcement learning (RL). However, prolonged CoT reasoning presents limitations, primarily verbose outputs due to excessive introspection. The reasoning process in these LLMs often appears to follow a trial-and-error methodology rather than a systematic, logical deduction. In contrast, tree-of-thoughts (ToT) offers a conceptually more advanced approach by modeling reasoning as an exploration within a tree structure. This reasoning structure facilitates the parallel generation and evaluation of multiple reasoning branches, allowing for the active identification, assessment, and pruning of unproductive paths. This process can potentially lead to improved performance and reduced token costs. Building upon the long CoT capability of LLMs, we introduce tree-of-thoughts RL (ToTRL), a novel on-policy RL framework with a rule-based reward. ToTRL is designed to guide LLMs in developing the parallel ToT strategy based on the sequential CoT strategy. Furthermore, we employ LLMs as players in a puzzle game during the ToTRL training process. Solving puzzle games inherently necessitates exploring interdependent choices and managing multiple constraints, which requires the construction and exploration of a thought tree, providing challenging tasks for cultivating the ToT reasoning capability. Our empirical evaluations demonstrate that our ToTQwen3-8B model, trained with our ToTRL, achieves significant improvement in performance and reasoning efficiency on complex reasoning tasks.

Rui Ming, Haoyuan Wu, Shoubo Hu et al.

Supervised fine-tuning (SFT) is the predominant method for adapting large language models (LLMs), yet it often struggles with generalization compared to reinforcement learning (RL). In this work, we posit that this performance disparity stems not just from the loss function, but from a more fundamental difference: SFT learns from a fixed, pre-collected dataset, whereas RL utilizes on-policy data sampled from the current policy. Building on this hypothesis, we introduce one-token rollout (OTR), a novel fine-tuning algorithm that guides SFT with the policy gradient method. OTR reframes the autoregressive learning process by treating each token generation as a single-step reinforcement learning trajectory. At each step, it performs a Monte Carlo ``rollout'' by sampling multiple candidate tokens from the current policy's distribution. The ground-truth token from the supervised data is then used to provide a reward signal to these samples. Guided by policy gradient, our algorithm repurposes static, off-policy supervised data into a dynamic, on-policy signal at the token level, capturing the generalization benefits of on-policy learning while bypassing the costly overhead of full sentence generation. Through extensive experiments on a diverse suite of challenging benchmarks spanning mathematical reasoning, code generation, and general domain reasoning, we demonstrate that OTR consistently outperforms standard SFT. Our findings establish OTR as a powerful and practical alternative for fine-tuning LLMs and provide compelling evidence that the on-policy nature of data is a critical driver of generalization, offering a promising new direction for fine-tuning LLMs.

Haoyuan Wu, Haisheng Zheng, Shoubo Hu et al.

Logic synthesis, a critical stage in electronic design automation (EDA), optimizes gate-level circuits to minimize power consumption and area occupancy in integrated circuits (ICs). Traditional logic synthesis tools rely on human-designed heuristics, often yielding suboptimal results. Although differentiable architecture search (DAS) has shown promise in generating circuits from truth tables, it faces challenges such as high computational complexity, convergence to local optima, and extensive hyperparameter tuning. Consequently, we propose a novel approach integrating conditional generative models with DAS for circuit generation. Our approach first introduces CircuitVQ, a circuit tokenizer trained based on our Circuit AutoEncoder We then develop CircuitAR, a masked autoregressive model leveraging CircuitVQ as the tokenizer. CircuitAR can generate preliminary circuit structures from truth tables, which guide DAS in producing functionally equivalent circuits. Notably, we observe the scalability and emergent capability in generating complex circuit structures of our CircuitAR models. Extensive experiments also show the superior performance of our method. This research bridges the gap between probabilistic generative models and precise circuit generation, offering a robust solution for logic synthesis.

Yunqi Wang, Furui Liu, Zhitang Chen et al.

Domain generalization aims to learn knowledge invariant across different distributions while semantically meaningful for downstream tasks from multiple source domains, to improve the model's generalization ability on unseen target domains. The fundamental objective is to understand the underlying "invariance" behind these observational distributions and such invariance has been shown to have a close connection to causality. While many existing approaches make use of the property that causal features are invariant across domains, we consider the causal invariance of the average causal effect of the features to the labels. This invariance regularizes our training approach in which interventions are performed on features to enforce stability of the causal prediction by the classifier across domains. Our work thus sheds some light on the domain generalization problem by introducing invariance of the mechanisms into the learning process. Experiments on several benchmark datasets demonstrate the performance of the proposed method against SOTAs.

Vahid Partovi Nia, Xinlin Li, Masoud Asgharian et al.

A probabilistic expert system emulates the decision-making ability of a human expert through a directional graphical model. The first step in building such systems is to understand data generation mechanism. To this end, one may try to decompose a multivariate distribution into product of several conditionals, and evolving a blackbox machine learning predictive models towards transparent cause-and-effect discovery. Most causal models assume a single homogeneous population, an assumption that may fail to hold in many applications. We show that when the homogeneity assumption is violated, causal models developed based on such assumption can fail to identify the correct causal direction. We propose an adjustment to a commonly used causal direction test statistic by using a $k$-means type clustering algorithm where both the labels and the number of components are estimated from the collected data to adjust the test statistic. Our simulation result show that the proposed adjustment significantly improves the performance of the causal direction test statistic for heterogeneous data. We study large sample behaviour of our proposed test statistic and demonstrate the application of the proposed method using real data.

Shoubo Hu, Kun Zhang, Zhitang Chen et al.

Domain generalization (DG) aims to incorporate knowledge from multiple source domains into a single model that could generalize well on unseen target domains. This problem is ubiquitous in practice since the distributions of the target data may rarely be identical to those of the source data. In this paper, we propose Multidomain Discriminant Analysis (MDA) to address DG of classification tasks in general situations. MDA learns a domain-invariant feature transformation that aims to achieve appealing properties, including a minimal divergence among domains within each class, a maximal separability among classes, and overall maximal compactness of all classes. Furthermore, we provide the bounds on excess risk and generalization error by learning theory analysis. Comprehensive experiments on synthetic and real benchmark datasets demonstrate the effectiveness of MDA.

Shoubo Hu, Zhitang Chen, Vahid Partovi Nia et al.

The inference of the causal relationship between a pair of observed variables is a fundamental problem in science, and most existing approaches are based on one single causal model. In practice, however, observations are often collected from multiple sources with heterogeneous causal models due to certain uncontrollable factors, which renders causal analysis results obtained by a single model skeptical. In this paper, we generalize the Additive Noise Model (ANM) to a mixture model, which consists of a finite number of ANMs, and provide the condition of its causal identifiability. To conduct model estimation, we propose Gaussian Process Partially Observable Model (GPPOM), and incorporate independence enforcement into it to learn latent parameter associated with each observation. Causal inference and clustering according to the underlying generating mechanisms of the mixture model are addressed in this work. Experiments on synthetic and real data demonstrate the effectiveness of our proposed approach.

Shoubo Hu, Zhitang Chen, Laiwan Chan

Although nonstationary data are more common in the real world, most existing causal discovery methods do not take nonstationarity into consideration. In this letter, we propose a kernel embedding-based approach, ENCI, for nonstationary causal model inference where data are collected from multiple domains with varying distributions. In ENCI, we transform the complicated relation of a cause-effect pair into a linear model of variables of which observations correspond to the kernel embeddings of the cause-and-effect distributions in different domains. In this way, we are able to estimate the causal direction by exploiting the causal asymmetry of the transformed linear model. Furthermore, we extend ENCI to causal graph discovery for multiple variables by transforming the relations among them into a linear nongaussian acyclic model. We show that by exploiting the nonstationarity of distributions, both cause-effect pairs and two kinds of causal graphs are identifiable under mild conditions. Experiments on synthetic and real-world data are conducted to justify the efficacy of ENCI over major existing methods.