Xinggao Liu

Hao Wang, Licheng Pan, Zhichao Chen et al.

Despite the success of reinforcement learning from human feedback (RLHF) in aligning language models, current reward modeling heavily relies on experimental feedback data collected from human annotators under controlled and costly conditions. In this work, we introduce observational reward modeling -- learning reward models with observational user feedback (e.g., clicks, copies, and upvotes) -- as a scalable and cost-effective alternative. We identify two fundamental challenges in this setting: (1) observational feedback is noisy due to annotation errors, which deviates it from true user preference; (2) observational feedback is biased by user preference, where users preferentially provide feedback on responses they feel strongly about, which creats a distribution shift between training and inference data. To address these challenges, we propose CausalRM, a causal-theoretic reward modeling framework that aims to learn unbiased reward models from observational feedback. To tackle challenge (1), CausalRM introduces a noise-aware surrogate loss term that is provably equivalent to the primal loss under noise-free conditions by explicitly modeling the annotation error generation process. To tackle challenge (2), CausalRM uses propensity scores -- the probability of a user providing feedback for a given response -- to reweight training samples, yielding a loss function that eliminates user preference bias. Extensive experiments across diverse LLM backbones and benchmark datasets validate that CausalRM effectively learns accurate reward signals from noisy and biased observational feedback and delivers substantial performance improvements on downstream RLHF tasks -- including a 49.2% gain on WildGuardMix and a 32.7% improvement on HarmBench. Code is available on our project website.

Hao Wang, Zhichao Chen, Yuan Shen et al.

Heterogeneous treatment effect (HTE) estimation from observational data poses significant challenges due to treatment selection bias. Existing methods address this bias by minimizing distribution discrepancies between treatment groups in latent space, focusing on global alignment. However, the fruitful aspect of local proximity, where similar units exhibit similar outcomes, is often overlooked. In this study, we propose Proximity-aware Counterfactual Regression (PCR) to exploit proximity for representation balancing within the HTE estimation context. Specifically, we introduce a local proximity preservation regularizer based on optimal transport to depict the local proximity in discrepancy calculation. Furthermore, to overcome the curse of dimensionality that renders the estimation of discrepancy ineffective, exacerbated by limited data availability for HTE estimation, we develop an informative subspace projector, which trades off minimal distance precision for improved sample complexity. Extensive experiments demonstrate that PCR accurately matches units across different treatment groups, effectively mitigates treatment selection bias, and significantly outperforms competitors. Code is available at https://anonymous.4open.science/status/ncr-B697.

Hao Wang, Zhichao Chen, Jiajun Fan et al.

As a basic research problem for building effective recommender systems, post-click conversion rate (CVR) estimation has long been plagued by sample selection bias and data sparsity issues. To address the data sparsity issue, prevalent methods based on entire space multi-task model leverage the sequential pattern of user actions, i.e. exposure $\rightarrow$ click $\rightarrow$ conversion to construct auxiliary learning tasks. However, they still fall short of guaranteeing the unbiasedness of CVR estimates. This paper theoretically demonstrates two defects of these entire space multi-task models: (1) inherent estimation bias (IEB) for CVR estimation, where the CVR estimate is inherently higher than the ground truth; (2) potential independence priority (PIP) for CTCVR estimation, where the causality from click to conversion might be overlooked. This paper further proposes a principled method named entire space counterfactual multi-task model (ESCM$^2$), which employs a counterfactual risk minimizer to handle both IEB and PIP issues at once. To demonstrate the effectiveness of the proposed method, this paper explores its parameter tuning in practice, derives its analytic properties, and showcases its effectiveness in industrial CVR estimation, where ESCM$^2$ can effectively alleviate the intrinsic IEB and PIP issues and outperform baseline models.

Licheng Pan, Hao Wang, Zhichao Chen et al.

Multi-variate soft sensor seeks accurate estimation of multiple quality variables using measurable process variables, which have emerged as a key factor in improving the quality of industrial manufacturing. The current progress stays in some direct applications of multitask network architectures; however, there are two fundamental issues remain yet to be investigated with these approaches: (1) negative transfer, where sharing representations despite the difference of discriminate representations for different objectives degrades performance; (2) seesaw phenomenon, where the optimizer focuses on one dominant yet simple objective at the expense of others. In this study, we reformulate the multi-variate soft sensor to a multi-objective problem, to address both issues and advance state-of-the-art performance. To handle the negative transfer issue, we first propose an Objective-aware Mixture-of-Experts (OMoE) module, utilizing objective-specific and objective-shared experts for parameter sharing while maintaining the distinction between objectives. To address the seesaw phenomenon, we then propose a Pareto Objective Routing (POR) module, adjusting the weights of learning objectives dynamically to achieve the Pareto optimum, with solid theoretical supports. We further present a Task-aware Mixture-of-Experts framework for achieving the Pareto optimum (TMoE-P) in multi-variate soft sensor, which consists of a stacked OMoE module and a POR module. We illustrate the efficacy of TMoE-P with an open soft sensor benchmark, where TMoE-P effectively alleviates the negative transfer and seesaw issues and outperforms the baseline models.

Hao Wang, Zhiyu Wang, Yunlong Niu et al.

An accurate and explainable automatic monitoring system is critical for the safety of high efficiency energy conversion plants that operate under extreme working condition. Nonetheless, currently available data-driven monitoring systems often fall short in meeting the requirements for either high-accuracy or interpretability, which hinders their application in practice. To overcome this limitation, a data-driven approach, AttentionMixer, is proposed under a generalized message passing framework, with the goal of establishing an accurate and interpretable radiation monitoring framework for energy conversion plants. To improve the model accuracy, the first technical contribution involves the development of spatial and temporal adaptive message passing blocks, which enable the capture of spatial and temporal correlations, respectively; the two blocks are cascaded through a mixing operator. To enhance the model interpretability, the second technical contribution involves the implementation of a sparse message passing regularizer, which eliminates spurious and noisy message passing routes. The effectiveness of the AttentionMixer approach is validated through extensive evaluations on a monitoring benchmark collected from the national radiation monitoring network for nuclear power plants, resulting in enhanced monitoring accuracy and interpretability in practice.

Xinjie Chen, Biao Fu, Jing Wu et al.

Reinforcement learning with verifiable rewards (RLVR) enhances the reasoning of large language models (LLMs), but standard RLVR often depends on human-annotated answers or carefully curated reward specifications. In machine-checkable domains, label-free alternatives such as majority voting or LLM-as-a-judge remove annotation cost but can introduce false positives that destabilize training. We introduce JURY-RL, a label-free RLVR framework that decouples answer proposal from reward disposal: votes from model rollouts propose a candidate answer, and a formal verifier determines whether that candidate can receive positive reward. Concretely, only rollouts matching the plurality-voted answer are rewarded when that answer is successfully verified in Lean. When verification is inconclusive, we invoke ResZero (Residual-Zero), a fallback reward that discards the unverified plurality proposal and redistributes a zero-mean, variance-preserving signal over the residual answers. This design maintains a stable optimization gradient without reinforcing unverifiable consensus. Across three backbone models trained on mathematical data, JURY-RL consistently outperforms other label-free baselines on mathematical reasoning benchmarks and transfers competitively to code generation and general benchmarks. It attains pass@1 performance comparable to supervised ground-truth training, with superior generalization demonstrated by higher pass@k and response diversity.

Hao Wang, Licheng Pan, Zhichao Chen et al. · pku

Time series modeling presents unique challenges due to autocorrelation in both historical data and future sequences. While current research predominantly addresses autocorrelation within historical data, the correlations among future labels are often overlooked. Specifically, modern forecasting models primarily adhere to the Direct Forecast (DF) paradigm, generating multi-step forecasts independently and disregarding label autocorrelation over time. In this work, we demonstrate that the learning objective of DF is biased in the presence of label autocorrelation. To address this issue, we propose the Frequency-enhanced Direct Forecast (FreDF), which mitigates label autocorrelation by learning to forecast in the frequency domain, thereby reducing estimation bias. Our experiments show that FreDF significantly outperforms existing state-of-the-art methods and is compatible with a variety of forecast models. Code is available at https://github.com/Master-PLC/FreDF.

Xinjie Chen, Minpeng Liao, Guoxin Chen et al.

Reinforcement learning with verifiable rewards (RLVR) has recently advanced the reasoning capabilities of large language models (LLMs). While prior work has emphasized algorithmic design, data curation, and reward shaping, we investigate RLVR from a sample-centric perspective and introduce LPPO (Learning-Progress and Prefix-guided Optimization), a framework of progressive optimization techniques. Our work addresses a critical question: how to best leverage a small set of trusted, high-quality demonstrations, rather than simply scaling up data volume. First, motivated by how hints aid human problem-solving, we propose prefix-guided sampling, an online data augmentation method that incorporates partial solution prefixes from expert demonstrations to guide the policy, particularly for challenging instances. Second, inspired by how humans focus on important questions aligned with their current capabilities, we introduce learning-progress weighting, a dynamic strategy that adjusts each training sample's influence based on model progression. We estimate sample-level learning progress via an exponential moving average of per-sample pass rates, promoting samples that foster learning and de-emphasizing stagnant ones. Experiments on mathematical-reasoning benchmarks demonstrate that our methods outperform strong baselines, yielding faster convergence and a higher performance ceiling.

Hao Wang, Zhichao Chen, Licheng Pan et al.

Effective process monitoring is increasingly vital in industrial automation for ensuring operational safety, necessitating both high accuracy and efficiency. Although Transformers have demonstrated success in various fields, their canonical form based on the self-attention mechanism is inadequate for process monitoring due to two primary limitations: (1) the step-wise correlations captured by self-attention mechanism are difficult to capture discriminative patterns in monitoring logs due to the lacking semantics of each step, thus compromising accuracy; (2) the quadratic computational complexity of self-attention hampers efficiency. To address these issues, we propose DeepFilter, a Transformer-style framework for process monitoring. The core innovation is an efficient filtering layer that excel capturing long-term and periodic patterns with reduced complexity. Equipping with the global filtering layer, DeepFilter enhances both accuracy and efficiency, meeting the stringent demands of process monitoring. Experimental results on real-world process monitoring datasets validate DeepFilter's superiority in terms of accuracy and efficiency compared to existing state-of-the-art models.

Yuxin Huang, Hao Wang, Zhaoran Liu et al.

Accurate estimation of multiple quality variables is critical for building industrial soft sensor models, which have long been confronted with data efficiency and negative transfer issues. Methods sharing backbone parameters among tasks address the data efficiency issue; however, they still fail to mitigate the negative transfer problem. To address this issue, a balanced Mixture-of-Experts (BMoE) is proposed in this work, which consists of a multi-gate mixture of experts (MMoE) module and a task gradient balancing (TGB) module. The MoE module aims to portray task relationships, while the TGB module balances the gradients among tasks dynamically. Both of them cooperate to mitigate the negative transfer problem. Experiments on the typical sulfur recovery unit demonstrate that BMoE models task relationship and balances the training process effectively, and achieves better performance than baseline models significantly.

Abiodun Ayodeji, Wenhai Wang, Jianzhong Su et al.

A single unit (head) is the conventional input feature extractor in deep learning architectures trained on multivariate time series signals. The importance of the fixed-dimensional vector representation generated by the single-head network has been demonstrated for industrial machinery condition monitoring and predictive maintenance. However, processing heterogeneous sensor signals with a single-head may result in a model that cannot explicitly account for the diversity in time-varying multivariate inputs. This work extends the conventional single-head deep learning models to a more robust form by developing context-specific heads to independently capture the inherent pattern in each sensor reading. Using the turbofan aircraft engine benchmark dataset (CMAPSS), an extensive experiment is performed to verify the effectiveness and benefits of multi-head multilayer perceptron, recurrent networks, convolution network, the transformer-style stand-alone attention network, and their variants for remaining useful life estimation. Moreover, the effect of different attention mechanisms on the multi-head models is also evaluated. In addition, each architecture's relative advantage and computational overhead are analyzed. Results show that utilizing the attention layer is task-sensitive and model dependent, as it does not provide consistent improvement across the models investigated. The best model is further compared with five state-of-the-art models, and the comparison shows that a relatively simple multi-head architecture performs better than the state-of-the-art models. The results presented in this study demonstrate the importance of multi-head models and attention mechanisms to an improved understanding of the remaining useful life of industrial assets.

Yilin Liao, Hao Wang, Zhaoran Liu et al.

The effectiveness of shortcut/skip-connection has been widely verified, which inspires massive explorations on neural architecture design. This work attempts to find an effective way to design new network architectures. It is discovered that the main difference between network architectures can be reflected in their recursion formulas. Based on this, a methodology is proposed to design novel network architectures from the perspective of mathematical formulas. Afterwards, a case study is provided to generate an improved architecture based on ResNet. Furthermore, the new architecture is compared with ResNet and then tested on ResNet-based networks. Massive experiments are conducted on CIFAR and ImageNet, which witnesses the significant performance improvements provided by the architecture.