Distilling reasoning paths from teacher to student models via supervised fine-tuning (SFT) provides a shortcut for improving the reasoning ability of the smaller Large Language Models (LLMs). However, the reasoning paths generated by teacher models often reflect only surface-level traces of their underlying authentic reasoning. Insights from cognitive neuroscience suggest that authentic reasoning involves a complex interweaving between meta-reasoning that selects the appropriate sub-problem from multiple candidates, and solving, which addresses the sub-problem. It means that authentic reasoning has implicit multi-branch structure. Supervised fine-tuning collapses this rich structure into a flat sequence of token prediction in teacher's reasoning path, which cannot distill this structure to student. To address this limitation, we propose RLKD, a reinforcement learning (RL)-based distillation framework guided by a novel Generative Structure Reward Model (GSRM). Our GSRM converts the reasoning path into multiple meta-reasoning-solving steps and gives the reward to measure the alignment between the reasoning structures of student and teacher. Our RLKD combines this reward with RL, enables the student LLM to internalize the teacher’s implicit multi-branch structure in authentic reasoning, rather than merely mimicking fixed teacher's output paths. Experiments show that RLKD, even when trained on only 0.1% of the data under an RL-only regime, surpasses the performance of standard SFT-RL pipelines and further unleashes the potential reasoning ability of the student LLM than SFT-based distillation.

As large language models (LLMs) are increasingly used as personalized user assistants, effectively adapting to users' evolving preferences is critical for delivering high-quality personalized responses. While user preferences are often stable in content, their relative strengths shift over time due to changing goals and contexts. Therefore, modeling these dynamic preference strengths can enable finer-grained personalization. However, current methods face two major challenges: (i) limited user feedback makes it difficult to estimate preference strengths accurately, and (ii) natural language ambiguity limits the controllability of preference-guided generation. To address these issues, we propose AdaPA-Agent, a LLM-agent personalization framework that models dynamic preference strengths via Adaptive Preference Arithmetic. First, instead of requiring additional user feedback, AdaPA-Agent employs an alignment-based strength estimation module to estimate the strength of user preferences from the existing user-agent interaction. Then, it guides controllable personalized generation by linearly combining next-token distributions, weighted by the estimated strengths of individual preferences. Experiments on two personalization tasks-conversational recommendation and personalized web interaction-demonstrate that AdaPA-Agent better aligning with users' changing intents, and has achieved over 18.9\% and 14.2\% improvements compared to ReAct, the widely-used agent framework.

MEMS gyroscopes play a critical role in inertial navigation and motion control applications but typically suffer from a fundamental trade-off between measurement range and noise performance. Existing hardware-based solutions aimed at mitigating this issue introduce additional complexity, cost, and scalability challenges. Deep-learning methods primarily focus on noise reduction and typically require precisely aligned ground-truth signals, making them difficult to deploy in practical scenarios and leaving the fundamental trade-off unresolved. To address these challenges, we introduce Mixture of Experts for MEMS Gyroscopes (MoE-Gyro), a novel self-supervised framework specifically designed for simultaneous over-range signal reconstruction and noise suppression. MoE-Gyro employs two experts: an Over‑Range Reconstruction Expert (ORE), featuring a Gaussian-Decay Attention mechanism for reconstructing saturated segments; and a Denoise Expert (DE), utilizing dual-branch complementary masking combined with FFT-guided augmentation for robust noise reduction. A lightweight gating module dynamically routes input segments to the appropriate expert. Furthermore, existing evaluation lack a comprehensive standard for assessing multi-dimensional signal enhancement. To bridge this gap, we introduce IMU Signal Enhancement Benchmark (ISEBench), an open-source benchmarking platform comprising the GyroPeak-100 dataset and a unified evaluation of IMU signal enhancement methods. We evaluate MoE-Gyro using our proposed ISEBench, demonstrating that our framework significantly extends the measurable range from ±450°/s to ±1500°/s, reduces Bias Instability by 98.4%, and achieves state-of-the-art performance, effectively addressing the long-standing trade-off in inertial sensing.Our code is available at: https://github.com/2002-Pan/Moe-Gyro

Constrained Reinforcement Learning (CRL) burgeons broad interest in recent years, which pursues maximizing long-term returns while constraining costs. Although CRL can be cast as a multi-objective optimization problem, it is still facing the key challenge that gradient-based Pareto optimization methods tend to stick to known Pareto-optimal solutions even when they yield poor returns (e.g., the safest self-driving car that never moves) or violate the constraints (e.g., the record-breaking racer that crashes the car). In this paper, we propose Gradient-adaptive Constrained Policy Optimization (GCPO for short), a novel Pareto optimization method for CRL with two adaptive gradient recalibration techniques. First, to find Pareto-optimal solutions with balanced performance over all targets, we propose gradient rebalancing which forces the agent to improve more on under-optimized objectives at every policy iteration. Second, to guarantee that the cost constraints are satisfied, we propose gradient perturbation that can temporarily sacrifice the returns for costs. Experiments on the SafetyGym benchmarks show that our method consistently outperforms previous CRL methods in reward while satisfying the constraints.

In the field of quantitative trading, it is common practice to transform raw historical stock data into indicative signals for the market trend. Such signals are called alpha factors. Alphas in formula forms are more interpretable and thus favored by practitioners concerned with risk. In practice, a set of formulaic alphas is often used together for better modeling precision, so we need to find synergistic formulaic alpha sets that work well together. However, most traditional alpha generators mine alphas one by one separately, overlooking the fact that the alphas would be combined later. In this paper, we propose a new alpha-mining framework that prioritizes mining a synergistic set of alphas, i.e., it directly uses the performance of the downstream combination model to optimize the alpha generator. Our framework also leverages the strong exploratory capabilities of reinforcement learning (RL) to better explore the vast search space of formulaic alphas. The contribution to the combination models' performance is assigned to be the return used in the RL process. This return drives the alpha generator to find better alphas that improve upon the current set. Experimental evaluations on real-world stock market data demonstrate both the effectiveness and the efficiency of our framework for stock trend forecasting. The investment simulation results show that our framework is able to achieve higher returns compared to previous approaches.

Optimal execution is a sequential decision-making problem for cost-saving in algorithmic trading. Studies have found that reinforcement learning (RL) can help decide the order-splitting sizes. However, a problem remains unsolved: how to place limit orders at appropriate limit prices? The key challenge lies in the ``continuous-discrete duality'' of the action space. On the one hand, the continuous action space using percentage changes in prices is preferred for generalization. On the other hand, the trader eventually needs to choose limit prices discretely due to the existence of the tick size, which requires specialization for every single stock with different characteristics (e.g., the liquidity and the price range). So we need continuous control for generalization and discrete control for specialization. To this end, we propose a hybrid RL method to combine the advantages of both of them. We first use a continuous control agent to scope an action subset, then deploy a fine-grained agent to choose a specific limit price. Extensive experiments show that our method has higher sample efficiency and better training stability than existing RL algorithms and significantly outperforms previous learning-based methods for order execution.

The delayed feedback problem is one of the imperative challenges in online advertising, which is caused by the highly diversified feedback delay of a conversion varying from a few minutes to several days. It is hard to design an appropriate online learning system under these non-identical delay for different types of ads and users. In this paper, we propose to tackle the delayed feedback problem in online advertising by "Following the Prophet" (FTP for short). The key insight is that, if the feedback came instantly for all the logged samples, we could get a model without delayed feedback, namely the "prophet". Although the prophet cannot be obtained during online learning, we show that we could predict the prophet's predictions by an aggregation policy on top of a set of multi-task predictions, where each task captures the feedback patterns of different periods. We propose the objective and optimization approach for the policy, and use the logged data to imitate the prophet. Extensive experiments on three real-world advertising datasets show that our method outperforms the previous state-of-the-art baselines.

The exploration/exploitation (E&E) dilemma lies at the core of interactive systems such as online advertising, for which contextual bandit algorithms have been proposed. Bayesian approaches provide guided exploration via uncertainty estimation, but the applicability is often limited due to over-simplified assumptions. Non-Bayesian bootstrap methods, on the other hand, can apply to complex problems by using deep reward models, but lack a clear guidance to the exploration behavior. It still remains largely unsolved to develop a practical method for complex deep contextual bandits.

Multi-view subspace learning aims to learn a shared representation from multiple sources or views of an entity. The learned representation enables reconstruction of common patterns of multi-view data, which helps dimensional reduction, exploratory data analysis, missing view completion, and various downstream tasks. However, existing methods often use simple structured approximations of the posterior of shared latent variables for the sake of computational efficiency. Such oversimplified models have a huge impact on the inference quality and can hurt the representation power. To this end, we propose a new method for multi-view subspace learning that achieves efficient Bayesian inference with strong representation power. Our method, coined CCA-Flow, bases on variational Canonical Correlation Analysis and models the inference network as an Inverse Autoregressive Flow (IAF). With the flow-based variational inference imposed on the latent variables, the posterior approximations can be arbitrarily complex and flexible, and the model can still be efficiently trained with stochastic gradient descent. Experiments on three benchmark multi-view datasets show that our model gives improved representations of shared latent variables and has superior performance against previous works.

It is a popular belief that model-based Reinforcement Learning (RL) is more sample efficient than model-free RL, but in practice, it is not always true due to overweighed model errors. In complex and noisy settings, model-based RL tends to have trouble using the model if it does not know when to trust the model. In this work, we find that better model usage can make a huge difference. We show theoretically that if the use of model-generated data is restricted to state-action pairs where the model error is small, the performance gap between model and real rollouts can be reduced. It motivates us to use model rollouts only when the model is confident about its predictions. We propose Masked Model-based Actor-Critic (M2AC), a novel policy optimization algorithm that maximizes a model-based lower-bound of the true value function. M2AC implements a masking mechanism based on the model's uncertainty estimation to decide whether the model should be used or not. Consequently, the new algorithm tends to give robust policy improvements. Experiments on continuous control benchmarks demonstrate that M2AC has strong performance even when using long model rollouts in very noisy environments, and significantly outperforms previous state-of-the-art methods.

A chatbot that converses like a human should be goal-oriented (i.e., be purposeful in conversation), which is beyond language generation. However, existing goal-oriented dialogue systems often heavily rely on cumbersome hand-crafted rules or costly labelled datasets, which limits the applicability. In this paper, we propose Goal-oriented Chatbots (GoChat), a framework for end-to-end training the chatbot to maximize the long-term return from offline multi-turn dialogue datasets. Our framework utilizes hierarchical reinforcement learning (HRL), where the high-level policy determines some sub-goals to guide the conversation towards the final goal, and the low-level policy fulfills the sub-goals by generating the corresponding utterance for response. In our experiments conducted on a real-world dialogue dataset for anti-fraud in financial, our approach outperforms previous methods on both the quality of response generation as well as the success rate of accomplishing the goal.

It is often observed that the probabilistic predictions given by a machine learning model can disagree with averaged actual outcomes on specific subsets of data, which is also known as the issue of miscalibration. It is responsible for the unreliability of practical machine learning systems. For example, in online advertising, an ad can receive a click-through rate prediction of 0.1 over some population of users where its actual click rate is 0.15. In such cases, the probabilistic predictions have to be fixed before the system can be deployed.

Model-free reinforcement learning methods such as the Proximal Policy Optimization algorithm (PPO) have successfully applied in complex decision-making problems such as Atari games. However, these methods suffer from high variances and high sample complexity. On the other hand, model-based reinforcement learning methods that learn the transition dynamics are more sample efficient, but they often suffer from the bias of the transition estimation. How to make use of both model-based and model-free learning is a central problem in reinforcement learning.In this paper, we present a new technique to address the tradeoff between exploration and exploitation, which regards the difference between model-free and model-based estimations as a measure of exploration value. We apply this new technique to the PPO algorithm and arrive at a new policy optimization method, named Policy Optimization with Modelbased Explorations (POME). POME uses two components to predict the actions’ target values: a model-free one estimated by Monte-Carlo sampling and a model-based one which learns a transition model and predicts the value of the next state. POME adds the error of these two target estimations as the additional exploration value for each state-action pair, i.e, encourages the algorithm to explore the states with larger target errors which are hard to estimate. We compare POME with PPO on Atari 2600 games, and it shows that POME outperforms PPO on 33 games out of 49 games.

Ling Luo, Xiang Ao, Yan Song, Feiyang Pan, Min Yang, Qing He. Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP). 2019.

Click-through rate (CTR) prediction has been one of the most central problems in computational advertising. Lately, embedding techniques that produce low-dimensional representations of ad IDs drastically improve CTR prediction accuracies. However, such learning techniques are data demanding and work poorly on new ads with little logging data, which is known as the cold-start problem.

Decision making is a challenging task in online recommender systems. The decision maker often needs to choose a contextual item at each step from a set of candidates. Contextual bandit algorithms have been successfully deployed to such applications, for the trade-off between exploration and exploitation and the state-of-art performance on minimizing online costs. However, the applicability of existing contextual bandit methods is limited by the over-simplified assumptions of the problem, such as assuming a simple form of the reward function or assuming a static environment where the states are not affected by previous actions.

Sentiment analysis has played a significant role in financial applications in recent years. The informational and emotive aspects of news texts may affect the prices, volatilities, volume of trades, and even potential risks of financial subjects. Previous studies in this field mainly focused on identifying polarity~(e.g. positive or negative). However, as financial decisions broadly require justifications, only plausible polarity cannot provide enough evidence during the decision making processes of humanity. Hence an explainable solution is in urgent demand. In this paper, we present an interpretable neural net framework for financial sentiment analysis. First, we design a hierarchical model to learn the representation of a document from multiple granularities. In addition, we propose a query-driven attention mechanism to satisfy the unique characteristics of financial documents. With the domain specified questions provided by the financial analysts, we can discover different spotlights for queries from different aspects. We conduct extensive experiments on a real-world dataset. The results demonstrate that our framework can learn better representation of the document and unearth meaningful clues on replying different users? preferences. It also outperforms the state-of-the-art methods on sentiment prediction of financial documents.