Hua Zheng

Shali Jiang, Hua Zheng, Boyang Liu et al.

Knowledge distillation (KD) transfers a single scalar prediction from a large foundation model (FM) to compact vertical models (VMs), suffering from diminishing transfer ratio -- the fraction of FM improvement captured by the VM -- as a single scalar cannot convey the rich intermediate knowledge that larger FMs learn. To address this bottleneck, we propose LoopFM (Learning frOm HistOrical ReP*resentations of FM), a framework that opens a high-bandwidth transfer channel by structuring FM intermediate embeddings as input features (e.g., user history sequence) for downstream VMs, without requiring real-time FM inference at serving and architectural coupling between FM and VM. We provide a theoretical framework for LoopFM with a gain decomposition and transfer-ratio analysis. On three public benchmarks, LoopFM demonstrates strong AUC improvements (e.g., 6\%+ on TaobaoAd) and complementary knowledge transfer capability with KD. On industrial-scale systems (billions of examples, trillion-parameter FMs), LoopFM approximately doubles the knowledge transfer ratio on top of KD, delivering a +0.5\% conversion improvement in Y1H1, and a +1.03\% and +1.22\% conversion improvement from two individual launches respectively in Y1H2.

Md Sakib Ullah Sourav, Huidong Wang, Mohammad Sultan Mahmud et al.

In this research, we propose a complete set of approaches for identifying and extracting emotions from Bangla texts. We provide a Bangla emotion classifier for six classes: anger, disgust, fear, joy, sadness, and surprise, from Bangla words using transformer-based models, which exhibit phenomenal results in recent days, especially for high-resource languages. The Unified Bangla Multi-class Emotion Corpus (UBMEC) is used to assess the performance of our models. UBMEC is created by combining two previously released manually labeled datasets of Bangla comments on six emotion classes with fresh manually labeled Bangla comments created by us. The corpus dataset and code we used in this work are publicly available.

Wei Xie, Keqi Wang, Hua Zheng et al.

Driven by the critical needs of biomanufacturing 4.0, we introduce a probabilistic knowledge graph hybrid model characterizing the risk- and science-based understanding of bioprocess mechanisms. It can faithfully capture the important properties, including nonlinear reactions, partially observed state, and nonstationary dynamics. Given very limited real process observations, we derive a posterior distribution quantifying model estimation uncertainty. To avoid the evaluation of intractable likelihoods, Approximate Bayesian Computation sampling with Sequential Monte Carlo (ABC-SMC) is utilized to approximate the posterior distribution. Under high stochastic and model uncertainties, it is computationally expensive to match output trajectories. Therefore, we create a linear Gaussian dynamic Bayesian network (LG-DBN) auxiliary likelihood-based ABC-SMC approach. Through matching the summary statistics driven through LG-DBN likelihood that can capture critical interactions and variations, the proposed algorithm can accelerate hybrid model inference, support process monitoring, and facilitate mechanism learning and robust control.

Hua Zheng, Wei Xie

Built on our previous study on green simulation assisted policy gradient (GS-PG) focusing on trajectory-based reuse, in this paper, we consider infinite-horizon Markov Decision Processes and create a new importance sampling based policy gradient optimization approach to support dynamic decision making. The existing GS-PG method was designed to learn from complete episodes or process trajectories, which limits its applicability to low-data situations and flexible online process control. To overcome this limitation, the proposed approach can selectively reuse the most related partial trajectories, i.e., the reuse unit is based on per-step or per-decision historical observations. In specific, we create a mixture likelihood ratio (MLR) based policy gradient optimization that can leverage the information from historical state-action transitions generated under different behavioral policies. The proposed variance reduction experience replay (VRER) approach can intelligently select and reuse most relevant transition observations, improve the policy gradient estimation, and accelerate the learning of optimal policy. Our empirical study demonstrates that it can improve optimization convergence and enhance the performance of state-of-the-art policy optimization approaches such as actor-critic method and proximal policy optimizations.

Hua Zheng, Wei Xie, M. Ben Feng

For reinforcement learning on complex stochastic systems where many factors dynamically impact the output trajectories, it is desirable to effectively leverage the information from historical samples collected in previous iterations to accelerate policy optimization. Classical experience replay allows agents to remember by reusing historical observations. However, the uniform reuse strategy that treats all observations equally overlooks the relative importance of different samples. To overcome this limitation, we propose a general variance reduction based experience replay (VRER) framework that can selectively reuse the most relevant samples to improve policy gradient estimation. This selective mechanism can adaptively put more weight on past samples that are more likely to be generated by the current target distribution. Our theoretical and empirical studies show that the proposed VRER can accelerate the learning of optimal policy and enhance the performance of state-of-the-art policy optimization approaches.

Hua Zheng, Kuang-Hung Liu, Igor Fedorov et al.

Neural Architecture Search (NAS) has become a widely used tool for automating neural network design. While one-shot NAS methods have successfully reduced computational requirements, they often require extensive training. On the other hand, zero-shot NAS utilizes training-free proxies to evaluate a candidate architecture's test performance but has two limitations: (1) inability to use the information gained as a network improves with training and (2) unreliable performance, particularly in complex domains like RecSys, due to the multi-modal data inputs and complex architecture configurations. To synthesize the benefits of both methods, we introduce a "sub-one-shot" paradigm that serves as a bridge between zero-shot and one-shot NAS. In sub-one-shot NAS, the supernet is trained using only a small subset of the training data, a phase we refer to as "warm-up." Within this framework, we present SiGeo, a proxy founded on a novel theoretical framework that connects the supernet warm-up with the efficacy of the proxy. Extensive experiments have shown that SiGeo, with the benefit of warm-up, consistently outperforms state-of-the-art NAS proxies on various established NAS benchmarks. When a supernet is warmed up, it can achieve comparable performance to weight-sharing one-shot NAS methods, but with a significant reduction ($\sim 60$\%) in computational costs.

Hua Zheng, Wei Xie, M. Ben Feng et al.

Effective reinforcement learning (RL) for complex stochastic systems requires leveraging historical data collected in previous iterations to accelerate policy optimization. Classical experience replay treats all past observations uniformly and fails to account for their varying contributions to learning. To overcome this limitation, we propose Variance Reduction Experience Replay (VRER), a principled framework that selectively reuses informative samples to reduce variance in policy gradient estimation. VRER is algorithm-agnostic and integrates seamlessly with existing policy optimization methods, forming the basis of our sample-efficient off-policy algorithm, Policy Gradient with VRER (PG-VRER). Motivated by the lack of rigorous theoretical analysis of experience replay, we develop a novel framework that explicitly captures dependencies introduced by Markovian dynamics and behavior-policy interactions. Using this framework, we establish finite-time convergence guarantees for PG-VRER and reveal a fundamental bias-variance trade-off: reusing older experience increases bias but simultaneously reduces gradient variance. Extensive empirical experiments demonstrate that VRER consistently accelerates policy learning and improves performance over state-of-the-art policy optimization algorithms.

Mingfu Liang, Xi Liu, Rong Jin et al.

Ads recommendation is a prominent service of online advertising systems and has been actively studied. Recent studies indicate that scaling-up and advanced design of the recommendation model can bring significant performance improvement. However, with a larger model scale, such prior studies have a significantly increasing gap from industry as they often neglect two fundamental challenges in industrial-scale applications. First, training and inference budgets are restricted for the model to be served, exceeding which may incur latency and impair user experience. Second, large-volume data arrive in a streaming mode with data distributions dynamically shifting, as new users/ads join and existing users/ads leave the system. We propose the External Large Foundation Model (ExFM) framework to address the overlooked challenges. Specifically, we develop external distillation and a data augmentation system (DAS) to control the computational cost of training/inference while maintaining high performance. We design the teacher in a way like a foundation model (FM) that can serve multiple students as vertical models (VMs) to amortize its building cost. We propose Auxiliary Head and Student Adapter to mitigate the data distribution gap between FM and VMs caused by the streaming data issue. Comprehensive experiments on internal industrial-scale applications and public datasets demonstrate significant performance gain by ExFM.

Chao Jiang, Dugang Liu, Cheng Wen et al.

Large language models have transformed AI-assisted software engineering, but current research remains biased toward high-resource languages such as Python, with weaker performance in languages like Rust and OCaml. Since real-world systems are inherently polyglot, robust multilingual code intelligence is crucial. This survey focuses on two key tasks: multilingual code generation from shared natural-language requirements, and multilingual code translation that preserves semantics across languages. It reviews representative methods, benchmarks, and evaluation metrics, and highlights challenges and opportunities for trustworthy cross-language generalization.

Xiangxiang Zhang, Jingxuan Wei, Donghong Zhong et al.

Existing Vision-Language Models often struggle with complex, multi-question reasoning tasks where partial correctness is crucial for effective learning. Traditional reward mechanisms, which provide a single binary score for an entire response, are too coarse to guide models through intricate problems with multiple sub-parts. To address this, we introduce StructVRM, a method that aligns multimodal reasoning with Structured and Verifiable Reward Models. At its core is a model-based verifier trained to provide fine-grained, sub-question-level feedback, assessing semantic and mathematical equivalence rather than relying on rigid string matching. This allows for nuanced, partial credit scoring in previously intractable problem formats. Extensive experiments demonstrate the effectiveness of StructVRM. Our trained model, Seed-StructVRM, achieves state-of-the-art performance on six out of twelve public multimodal benchmarks and our newly curated, high-difficulty STEM-Bench. The success of StructVRM validates that training with structured, verifiable rewards is a highly effective approach for advancing the capabilities of multimodal models in complex, real-world reasoning domains.

Hua Zheng, Wei Xie, Ilya O. Ryzhov et al.

This paper presents a novel methodological framework, called the Actor-Simulator, that incorporates the calibration of digital twins into model-based reinforcement learning for more effective control of stochastic systems with complex nonlinear dynamics. Traditional model-based control often relies on restrictive structural assumptions (such as linear state transitions) and fails to account for parameter uncertainty in the model. These issues become particularly critical in industries such as biopharmaceutical manufacturing, where process dynamics are complex and not fully known, and only a limited amount of data is available. Our approach jointly calibrates the digital twin and searches for an optimal control policy, thus accounting for and reducing model error. We balance exploration and exploitation by using policy performance as a guide for data collection. This dual-component approach provably converges to the optimal policy, and outperforms existing methods in extensive numerical experiments based on the biopharmaceutical manufacturing domain.

Fuqiang Cheng, Wei Xie, Hua Zheng

Biomanufacturing innovation relies on an efficient Design of Experiments (DoEs) to optimize processes and product quality. Traditional DoE methods, ignoring the underlying bioprocessing mechanisms, often suffer from a lack of interpretability and sample efficiency. This limitation motivates us to create a new optimal learning approach for digital twin model calibration. In this study, we consider the cell culture process multi-scale mechanistic model, also known as Biological System-of-Systems (Bio-SoS). This model with a modular design, composed of sub-models, allows us to integrate data across various production processes. To calibrate the Bio-SoS digital twin, we evaluate the mean squared error of model prediction and develop a computational approach to quantify the impact of parameter estimation error of individual sub-models on the prediction accuracy of digital twin, which can guide sample-efficient and interpretable DoEs.

Hua Zheng, Wei Xie, Keqi Wang et al.

Driven by the key challenges of cell therapy manufacturing, including high complexity, high uncertainty, and very limited process observations, we propose a hybrid model-based reinforcement learning (RL) to efficiently guide process control. We first create a probabilistic knowledge graph (KG) hybrid model characterizing the risk- and science-based understanding of biomanufacturing process mechanisms and quantifying inherent stochasticity, e.g., batch-to-batch variation. It can capture the key features, including nonlinear reactions, nonstationary dynamics, and partially observed state. This hybrid model can leverage existing mechanistic models and facilitate learning from heterogeneous process data. A computational sampling approach is used to generate posterior samples quantifying model uncertainty. Then, we introduce hybrid model-based Bayesian RL, accounting for both inherent stochasticity and model uncertainty, to guide optimal, robust, and interpretable dynamic decision making. Cell therapy manufacturing examples are used to empirically demonstrate that the proposed framework can outperform the classical deterministic mechanistic model assisted process optimization.

Qing Lyu, Hua Zheng, Daoxin Li et al.

Recursive noun phrases (NPs) have interesting semantic properties. For example, "my favorite new movie" is not necessarily my favorite movie, whereas "my new favorite movie" is. This is common sense to humans, yet it is unknown whether language models have such knowledge. We introduce the Recursive Noun Phrase Challenge (RNPC), a dataset of three textual inference tasks involving textual entailment and event plausibility comparison, precisely targeting the understanding of recursive NPs. When evaluated on RNPC, state-of-the-art Transformer models only perform around chance. Still, we show that such knowledge is learnable with appropriate data. We further probe the models for relevant linguistic features that can be learned from our tasks, including modifier semantic category and modifier scope. Finally, models trained on RNPC achieve strong zero-shot performance on an extrinsic Harm Detection evaluation task, showing the usefulness of the understanding of recursive NPs in downstream applications.

Hua Zheng, Wei Xie, M. Ben Feng

Experience replay is a core ingredient of modern deep reinforcement learning, yet its benefits in policy optimization are poorly understood beyond empirical heuristics. This paper develops a novel theoretical framework for experience replay in modern policy gradient methods, where two sources of dependence fundamentally complicate analysis: Markovian correlations along trajectories and policy drift across optimization iterations. We introduce a new proof technique based on auxiliary Markov chains and lag-based decoupling that makes these dependencies tractable. Within this framework, we derive finite-time bias bounds for policy-gradient estimators under replay, identifying how bias scales with the cumulative policy update, the mixing time of the underlying dynamics, and the age of buffered data, thereby formalizing the practitioner's rule of avoiding overly stale replay. We further provide a correlation-aware variance decomposition showing how sample dependence governs gradient variance from replay and when replay is beneficial. Building on these characterizations, we establish the finite-time convergence guarantees for experience-replay-based policy optimization, explicitly quantifying how buffer size, sample correlation, and mixing jointly determine the convergence rate and revealing an inherent bias-variance trade-off: larger buffers can reduce variance by averaging less correlated samples but can increase bias as data become stale. These results offer a principled guide for buffer sizing and replay schedules, bridging prior empirical findings with quantitative theory.

Hua Zheng, Jiahao Zhu, Wei Xie et al.

Patients with severe Coronavirus disease 19 (COVID-19) typically require supplemental oxygen as an essential treatment. We developed a machine learning algorithm, based on a deep Reinforcement Learning (RL), for continuous management of oxygen flow rate for critical ill patients under intensive care, which can identify the optimal personalized oxygen flow rate with strong potentials to reduce mortality rate relative to the current clinical practice. Basically, we modeled the oxygen flow trajectory of COVID-19 patients and their health outcomes as a Markov decision process. Based on individual patient characteristics and health status, a reinforcement learning based oxygen control policy is learned and real-time recommends the oxygen flow rate to reduce the mortality rate. We assessed the performance of proposed methods through cross validation by using a retrospective cohort of 1,372 critically ill patients with COVID-19 from New York University Langone Health ambulatory care with electronic health records from April 2020 to January 2021. The mean mortality rate under the RL algorithm is lower than standard of care by 2.57% (95% CI: 2.08- 3.06) reduction (P<0.001) from 7.94% under the standard of care to 5.37 % under our algorithm and the averaged recommended oxygen flow rate is 1.28 L/min (95% CI: 1.14-1.42) lower than the rate actually delivered to patients. Thus, the RL algorithm could potentially lead to better intensive care treatment that can reduce mortality rate, while saving the oxygen scarce resources. It can reduce the oxygen shortage issue and improve public health during the COVID-19 pandemic.

Hua Zheng, Wei Xie, Ilya O. Ryzhov et al.

Biopharmaceutical manufacturing is a rapidly growing industry with impact in virtually all branches of medicines. Biomanufacturing processes require close monitoring and control, in the presence of complex bioprocess dynamics with many interdependent factors, as well as extremely limited data due to the high cost of experiments as well as the novelty of personalized bio-drugs. We develop a novel model-based reinforcement learning framework that can achieve human-level control in low-data environments. The model uses a dynamic Bayesian network to capture causal interdependencies between factors and predict how the effects of different inputs propagate through the pathways of the bioprocess mechanisms. This enables the design of process control policies that are both interpretable and robust against model risk. We present a computationally efficient, provably convergence stochastic gradient method for optimizing such policies. Validation is conducted on a realistic application with a multi-dimensional, continuous state variable.

Damai Dai, Hua Zheng, Zhifang Sui et al.

Conventional Machine Reading Comprehension (MRC) has been well-addressed by pattern matching, but the ability of commonsense reasoning remains a gap between humans and machines. Previous methods tackle this problem by enriching word representations via pre-trained Knowledge Graph Embeddings (KGE). However, they make limited use of a large number of connections between nodes in Knowledge Graphs (KG), which could be pivotal cues to build the commonsense reasoning chains. In this paper, we propose a Plug-and-play module to IncorporatE Connection information for commonsEnse Reasoning (PIECER). Beyond enriching word representations with knowledge embeddings, PIECER constructs a joint query-passage graph to explicitly guide commonsense reasoning by the knowledge-oriented connections between words. Further, PIECER has high generalizability since it can be plugged into suitable positions in any MRC model. Experimental results on ReCoRD, a large-scale public MRC dataset requiring commonsense reasoning, show that PIECER introduces stable performance improvements for four representative base MRC models, especially in low-resource settings.

Hua Zheng, Ilya O. Ryzhov, Wei Xie et al.

Comorbid chronic conditions are common among people with type 2 diabetes. We developed an Artificial Intelligence algorithm, based on Reinforcement Learning (RL), for personalized diabetes and multi-morbidity management with strong potential to improve health outcomes relative to current clinical practice. In this paper, we modeled glycemia, blood pressure and cardiovascular disease (CVD) risk as health outcomes using a retrospective cohort of 16,665 patients with type 2 diabetes from New York University Langone Health ambulatory care electronic health records in 2009 to 2017. We trained a RL prescription algorithm that recommends a treatment regimen optimizing patients' cumulative health outcomes using their individual characteristics and medical history at each encounter. The RL recommendations were evaluated on an independent subset of patients. The results demonstrate that the proposed personalized reinforcement learning prescriptive framework for type 2 diabetes yielded high concordance with clinicians' prescriptions and substantial improvements in glycemia, blood pressure, cardiovascular disease risk outcomes.

Damai Dai, Hua Zheng, Fuli Luo et al.

Conventional Knowledge Graph Completion (KGC) assumes that all test entities appear during training. However, in real-world scenarios, Knowledge Graphs (KG) evolve fast with out-of-knowledge-graph (OOKG) entities added frequently, and we need to represent these entities efficiently. Most existing Knowledge Graph Embedding (KGE) methods cannot represent OOKG entities without costly retraining on the whole KG. To enhance efficiency, we propose a simple and effective method that inductively represents OOKG entities by their optimal estimation under translational assumptions. Given pretrained embeddings of the in-knowledge-graph (IKG) entities, our method needs no additional learning. Experimental results show that our method outperforms the state-of-the-art methods with higher efficiency on two KGC tasks with OOKG entities.

Hua Zheng, Wei Xie, Mingbin Ben Feng

Biopharmaceutical manufacturing faces critical challenges, including complexity, high variability, lengthy lead time, and limited historical data and knowledge of the underlying system stochastic process. To address these challenges, we propose a green simulation assisted model-based reinforcement learning to support process online learning and guide dynamic decision making. Basically, the process model risk is quantified by the posterior distribution. At any given policy, we predict the expected system response with prediction risk accounting for both inherent stochastic uncertainty and model risk. Then, we propose green simulation assisted reinforcement learning and derive the mixture proposal distribution of decision process and likelihood ratio based metamodel for the policy gradient, which can selectively reuse process trajectory outputs collected from previous experiments to increase the simulation data-efficiency, improve the policy gradient estimation accuracy, and speed up the search for the optimal policy. Our numerical study indicates that the proposed approach demonstrates the promising performance.

Sumitra Ganesh, Nelson Vadori, Mengda Xu et al.

Market makers play an important role in providing liquidity to markets by continuously quoting prices at which they are willing to buy and sell, and managing inventory risk. In this paper, we build a multi-agent simulation of a dealer market and demonstrate that it can be used to understand the behavior of a reinforcement learning (RL) based market maker agent. We use the simulator to train an RL-based market maker agent with different competitive scenarios, reward formulations and market price trends (drifts). We show that the reinforcement learning agent is able to learn about its competitor's pricing policy; it also learns to manage inventory by smartly selecting asymmetric prices on the buy and sell sides (skewing), and maintaining a positive (or negative) inventory depending on whether the market price drift is positive (or negative). Finally, we propose and test reward formulations for creating risk averse RL-based market maker agents.

Wei Xie, Yuan Yi, Hua Zheng

To integrate strategic, tactical and operational decisions, the two-stage optimization has been widely used to guide dynamic decision making. In this paper, we study the two-stage stochastic programming for complex systems with unknown response estimated by simulation. We introduce the global-local metamodel assisted two-stage optimization via simulation that can efficiently employ the simulation resource to iteratively solve for the optimal first- and second-stage decisions. Specifically, at each visited first-stage decision, we develop a local metamodel to simultaneously solve a set of scenario-based second-stage optimization problems, which also allows us to estimate the optimality gap. Then, we construct a global metamodel accounting for the errors induced by: (1) using a finite number of scenarios to approximate the expected future cost occurring in the planning horizon, (2) second-stage optimality gap, and (3) finite visited first-stage decisions. Assisted by the global-local metamodel, we propose a new simulation optimization approach that can efficiently and iteratively search for the optimal first- and second-stage decisions. Our framework can guarantee the convergence of optimal solution for the discrete two-stage optimization with unknown objective, and the empirical study indicates that it achieves substantial efficiency and accuracy.