arXiv subjects

Jianfei Zhang

Publications and source records attributed to Jianfei Zhang.

At least 19 recordsLinked to original sources

arXiv2026

Survival Analysis with Machine Learning for Predicting Li-ion Battery Remaining Useful Life

Battery degradation significantly impacts the reliability and efficiency of energy storage systems, particularly in electric vehicles and industrial applications. Predicting the remaining useful life (RUL) of lithium-ion batteries is crucial for optimizing maintenance schedules, reducing costs, and improving safety. Traditional RUL prediction methods often struggle with nonlinear degradation patterns and uncertainty quantification. To address these challenges, we propose a hybrid survival analysis framework integrating survival data reconstruction, survival model learning, and survival probability estimation. Our approach transforms battery voltage time series into time-to-failure data using path signatures. The multiple Cox-based survival models and machine-learning-based methods, such as DeepHit and MTLR, are learned to predict battery failure-free probabilities over time. Experiments conducted on the Toyota battery and NASA battery datasets demonstrate the effectiveness of our approach, achieving high time-dependent AUC and concordance index (C-Index) while maintaining a low integrated Brier score. The data and source codes are available to the public at https://github.com/okic-ca/rul

eess.SP↗

arXiv2026

When Model Merging Breaks Routing: Training-Free Calibration for MoE

Model merging has emerged as a cost-effective approach for consolidating the capabilities of multiple LLMs without retraining. However, existing merging techniques, largely based on linear parameter arithmetic or optimization, struggle when applied to Mixture-of-Experts (MoE) architectures. We identify a critical failure mode in MoE merging, termed routing breakdown, in which the merged router fails to dispatch tokens to suitable experts. Routing breakdown stems from the sensitivity of the non-linear softmax and discrete Top-k routing mechanisms to parameter perturbations from merging, a sensitivity further amplified by load-balancing constraints imposed during MoE pretraining. Because fine-tuned experts exhibit distinct specializations, even modest misrouting can cause severe performance degradation. To address this issue, we propose Hessian-Aware Router Calibration (HARC), a training-free framework that leverages second-order curvature information to realign the merged router. This approach admits a closed-form solution that can be efficiently solved using a matrix-free conjugate gradient method. Experiments on mathematical reasoning and code generation tasks show that HARC effectively mitigates routing breakdown across diverse MoE merging baselines and leads to substantial performance improvements. Our code is available at https://github.com/huangcb01/HARC.

cs.LG↗

arXiv2026

LongCat-Flash-Prover: Advancing Native Formal Reasoning via Agentic Tool-Integrated Reinforcement Learning

We introduce LongCat-Flash-Prover, a flagship 560-billion-parameter open-source Mixture-of- Experts (MoE) model that advances Native Formal Reasoning in Lean4 through agentic tool-integrated reasoning (TIR). We decompose the native formal reasoning task into three independent formal capabilities, i.e., auto-formalization, sketching, and proving. To facilitate these capabilities, we propose a Hybrid-Experts Iteration Framework to expand high-quality task trajectories, including generating a formal statement based on a given informal problem, producing a whole-proof directly from the statement, or a lemma-style sketch. During agentic RL, we present a Hierarchical Importance Sampling Policy Optimization (HisPO) algorithm, which aims to stabilize the MoE model training on such long-horizon tasks. It employs a gradient masking strategy that accounts for the policy staleness and the inherent train-inference engine discrepancies at both sequence and token levels. Additionally, we also incorporate theorem consistency and legality detection mechanisms to eliminate reward hacking issues. Extensive evaluations show that our LongCat-Flash-Prover sets a new state-of-the-art for open-weights models in both auto-formalization and theorem proving. Demonstrating remarkable sample efficiency, it achieves a 97.1% pass rate on MiniF2F-Test using only 72 inference budget per problem. On more challenging benchmarks, it solves 70.8% of ProverBench and 41.5% of PutnamBench with no more than 220 attempts per problem, significantly outperforming existing open-weights baselines.

cs.AI↗

arXiv2026

OPE: Overcoming Information Saturation in Parallel Thinking via Outline-Guided Path Exploration

Parallel thinking has emerged as a new paradigm for large reasoning models (LRMs) in tackling complex problems. Recent methods leverage Reinforcement Learning (RL) to enhance parallel thinking, aiming to address the limitations in computational resources and effectiveness encountered with supervised fine-tuning. However, most existing studies primarily focus on optimizing the aggregation phase, with limited attention to the path exploration stage. In this paper, we theoretically analyze the optimization of parallel thinking under the Reinforcement Learning with Verifiable Rewards (RLVR) setting, and identify that the mutual information bottleneck among exploration paths fundamentally restricts overall performance. To address this, we propose Outline-Guided Path Exploration (OPE), which explicitly partitions the solution space by generating diverse reasoning outlines prior to parallel path reasoning, thereby reducing information redundancy and improving the diversity of information captured across exploration paths. We implement OPE with an iterative RL strategy that optimizes outline planning and outline-guided reasoning independently. Extensive experiments across multiple challenging mathematical benchmarks demonstrate that OPE effectively improves reasoning performance in different aggregation strategies, enabling LRMs to more reliably discover correct solutions.

cs.AI↗

arXiv2026

LongCat-Flash-Thinking-2601 Technical Report

We introduce LongCat-Flash-Thinking-2601, a 560-billion-parameter open-source Mixture-of-Experts (MoE) reasoning model with superior agentic reasoning capability. LongCat-Flash-Thinking-2601 achieves state-of-the-art performance among open-source models on a wide range of agentic benchmarks, including agentic search, agentic tool use, and tool-integrated reasoning. Beyond benchmark performance, the model demonstrates strong generalization to complex tool interactions and robust behavior under noisy real-world environments. Its advanced capability stems from a unified training framework that combines domain-parallel expert training with subsequent fusion, together with an end-to-end co-design of data construction, environments, algorithms, and infrastructure spanning from pre-training to post-training. In particular, the model's strong generalization capability in complex tool-use are driven by our in-depth exploration of environment scaling and principled task construction. To optimize long-tailed, skewed generation and multi-turn agentic interactions, and to enable stable training across over 10,000 environments spanning more than 20 domains, we systematically extend our asynchronous reinforcement learning framework, DORA, for stable and efficient large-scale multi-environment training. Furthermore, recognizing that real-world tasks are inherently noisy, we conduct a systematic analysis and decomposition of real-world noise patterns, and design targeted training procedures to explicitly incorporate such imperfections into the training process, resulting in improved robustness for real-world applications. To further enhance performance on complex reasoning tasks, we introduce a Heavy Thinking mode that enables effective test-time scaling by jointly expanding reasoning depth and width through intensive parallel thinking.

cs.AI↗

arXiv2025

Introducing LongCat-Flash-Thinking: A Technical Report

We present LongCat-Flash-Thinking, an efficient 560-billion-parameter open-source Mixture-of-Experts (MoE) reasoning model. Its advanced capabilities are cultivated through a meticulously crafted training process, beginning with long Chain-of-Thought (CoT) data cold-start and culminating in large-scale Reinforcement Learning (RL). We first employ a well-designed cold-start training strategy, which significantly enhances the reasoning potential and equips the model with specialized skills in both formal and agentic reasoning. Then, a core innovation is our domain-parallel training scheme, which decouples optimization across distinct domains (e.g., STEM, Code, Agentic) and subsequently fuses the resulting expert models into a single, nearly Pareto-optimal model. This entire process is powered by our Dynamic ORchestration for Asynchronous rollout (DORA) system, a large-scale RL framework that delivers a greater than threefold training speedup over synchronous methods on tens of thousands of accelerators. As a result, LongCat-Flash-Thinking achieves state-of-the-art performance among open-source models on a suite of complex reasoning tasks. The model exhibits exceptional efficiency in agentic reasoning, reducing average token consumption by 64.5% (from 19, 653 to 6, 965) on AIME-25, without degrading task accuracy. We release LongCat-Flash-Thinking to promote further advances in reasoning systems and agentic AI research.

cs.AI↗

arXiv2025

Autoformalizer with Tool Feedback

Autoformalization addresses the scarcity of data for Automated Theorem Proving (ATP) by translating mathematical problems from natural language into formal statements. Efforts in recent work shift from directly prompting large language models to training an end-to-end formalizer model from scratch, achieving remarkable advancements. However, existing formalizer still struggles to consistently generate valid statements that meet syntactic validity and semantic consistency. To address this issue, we propose the Autoformalizer with Tool Feedback (ATF), a novel approach that incorporates syntactic and consistency information as tools into the formalization process. By integrating Lean 4 compilers for syntax corrections and employing a multi-LLMs-as-judge approach for consistency validation, the model is able to adaptively refine generated statements according to the tool feedback, enhancing both syntactic validity and semantic consistency. The training of ATF involves a cold-start phase on synthetic tool-calling data, an expert iteration phase to improve formalization capabilities, and Direct Preference Optimization to alleviate ineffective revisions. Experimental results show that ATF markedly outperforms a range of baseline formalizer models, with its superior performance further validated by human evaluations. Subsequent analysis reveals that ATF demonstrates excellent inference scaling properties. Moreover, we open-source Numina-ATF, a dataset containing 750K synthetic formal statements to facilitate advancements in autoformalization and ATP research.

cs.AI↗

arXiv2025

Selecting Demonstrations for Many-Shot In-Context Learning via Gradient Matching

In-Context Learning (ICL) empowers Large Language Models (LLMs) for rapid task adaptation without Fine-Tuning (FT), but its reliance on demonstration selection remains a critical challenge. While many-shot ICL shows promising performance through scaled demonstrations, the selection method for many-shot demonstrations remains limited to random selection in existing work. Since the conventional instance-level retrieval is not suitable for many-shot scenarios, we hypothesize that the data requirements for in-context learning and fine-tuning are analogous. To this end, we introduce a novel gradient matching approach that selects demonstrations by aligning fine-tuning gradients between the entire training set of the target task and the selected examples, so as to approach the learning effect on the entire training set within the selected examples. Through gradient matching on relatively small models, e.g., Qwen2.5-3B or Llama3-8B, our method consistently outperforms random selection on larger LLMs from 4-shot to 128-shot scenarios across 9 diverse datasets. For instance, it surpasses random selection by 4% on Qwen2.5-72B and Llama3-70B, and by around 2% on 5 closed-source LLMs. This work unlocks more reliable and effective many-shot ICL, paving the way for its broader application.

cs.CL↗

arXiv2025

Agent-based Liquidity Risk Modelling for Financial Markets

In this paper, we describe a novel agent-based approach for modelling the transaction cost of buying or selling an asset in financial markets, e.g., to liquidate a large position as a result of a margin call to meet financial obligations. The simple act of buying or selling in the market causes a price impact and there is a cost described as liquidity risk. For example, when selling a large order, there is market slippage -- each successive trade will execute at the same or worse price. When the market adjusts to the new information revealed by the execution of such a large order, we observe in the data a permanent price impact that can be attributed to the change in the fundamental value as market participants reassess the value of the asset. In our ABM model, we introduce a novel mechanism where traders assume orderflow is informed and each trade reveals some information about the value of the asset, and traders update their belief of the fundamental value for every trade. The result is emergent, realistic price impact without oversimplifying the problem as most stylised models do, but within a realistic framework that models the exchange with its protocols, its limit orderbook and its auction mechanism and that can calculate the transaction cost of any execution strategy without limitation. Our stochastic ABM model calculates the costs and uncertainties of buying and selling in a market by running Monte-Carlo simulations, for a better understanding of liquidity risk and can be used to optimise for optimal execution under liquidity risk. We demonstrate its practical application in the real world by calculating the liquidity risk for the Hang-Seng Futures Index.

q-fin.TR↗

arXiv2025

AAKT: Enhancing Knowledge Tracing with Alternate Autoregressive Modeling

Knowledge Tracing (KT) aims to predict students' future performances based on their former exercises and additional information in educational settings. KT has received significant attention since it facilitates personalized experiences in educational situations. Simultaneously, the autoregressive modeling on the sequence of former exercises has been proven effective for this task. One of the primary challenges in autoregressive modeling for Knowledge Tracing is effectively representing the anterior (pre-response) and posterior (post-response) states of learners across exercises. Existing methods often employ complex model architectures to update learner states using question and response records. In this study, we propose a novel perspective on knowledge tracing task by treating it as a generative process, consistent with the principles of autoregressive models. We demonstrate that knowledge states can be directly represented through autoregressive encodings on a question-response alternate sequence, where model generate the most probable representation in hidden state space by analyzing history interactions. This approach underpins our framework, termed Alternate Autoregressive Knowledge Tracing (AAKT). Additionally, we incorporate supplementary educational information, such as question-related skills, into our framework through an auxiliary task, and include extra exercise details, like response time, as additional inputs. Our proposed framework is implemented using advanced autoregressive technologies from Natural Language Generation (NLG) for both training and prediction. Empirical evaluations on four real-world KT datasets indicate that AAKT consistently outperforms all baseline models in terms of AUC, ACC, and RMSE. Furthermore, extensive ablation studies and visualized analysis validate the effectiveness of key components in AAKT.

cs.AI↗

arXiv2024

Disentangling Preference Representation and Text Generation for Efficient Individual Preference Alignment

Aligning Large Language Models (LLMs) with general human preferences has been proved crucial in improving the interaction quality between LLMs and human. However, human values are inherently diverse among different individuals, making it insufficient to align LLMs solely with general preferences. To address this, personalizing LLMs according to individual feedback emerges as a promising solution. Nonetheless, this approach presents challenges in terms of the efficiency of alignment algorithms. In this work, we introduce a flexible paradigm for individual preference alignment. Our method fundamentally improves efficiency by disentangling preference representation from text generation in LLMs. We validate our approach across multiple text generation tasks and demonstrate that it can produce aligned quality as well as or better than PEFT-based methods, while reducing additional training time for each new individual preference by $80\%$ to $90\%$ in comparison with them.

cs.CL↗

arXiv2024

Leveraging Estimated Transferability Over Human Intuition for Model Selection in Text Ranking

Text ranking has witnessed significant advancements, attributed to the utilization of dual-encoder enhanced by Pre-trained Language Models (PLMs). Given the proliferation of available PLMs, selecting the most effective one for a given dataset has become a non-trivial challenge. As a promising alternative to human intuition and brute-force fine-tuning, Transferability Estimation (TE) has emerged as an effective approach to model selection. However, current TE methods are primarily designed for classification tasks, and their estimated transferability may not align well with the objectives of text ranking. To address this challenge, we propose to compute the expected rank as transferability, explicitly reflecting the model's ranking capability. Furthermore, to mitigate anisotropy and incorporate training dynamics, we adaptively scale isotropic sentence embeddings to yield an accurate expected rank score. Our resulting method, Adaptive Ranking Transferability (AiRTran), can effectively capture subtle differences between models. On challenging model selection scenarios across various text ranking datasets, it demonstrates significant improvements over previous classification-oriented TE methods, human intuition, and ChatGPT with minor time consumption.

cs.AI↗

arXiv2024

ProCQA: A Large-scale Community-based Programming Question Answering Dataset for Code Search

Retrieval-based code question answering seeks to match user queries in natural language to relevant code snippets. Previous approaches typically rely on pretraining models using crafted bi-modal and uni-modal datasets to align text and code representations. In this paper, we introduce ProCQA, a large-scale programming question answering dataset extracted from the StackOverflow community, offering naturally structured mixed-modal QA pairs. To validate its effectiveness, we propose a modality-agnostic contrastive pre-training approach to improve the alignment of text and code representations of current code language models. Compared to previous models that primarily employ bimodal and unimodal pairs extracted from CodeSearchNet for pre-training, our model exhibits significant performance improvements across a wide range of code retrieval benchmarks.

cs.CL↗

arXiv2023

Deep Calibration of Market Simulations using Neural Density Estimators and Embedding Networks

The ability to construct a realistic simulator of financial exchanges, including reproducing the dynamics of the limit order book, can give insight into many counterfactual scenarios, such as a flash crash, a margin call, or changes in macroeconomic outlook. In recent years, agent-based models have been developed that reproduce many features of an exchange, as summarised by a set of stylised facts and statistics. However, the ability to calibrate simulators to a specific period of trading remains an open challenge. In this work, we develop a novel approach to the calibration of market simulators by leveraging recent advances in deep learning, specifically using neural density estimators and embedding networks. We demonstrate that our approach is able to correctly identify high probability parameter sets, both when applied to synthetic and historical data, and without reliance on manually selected or weighted ensembles of stylised facts.

cs.LG↗

arXiv2023

Noise-induced dynamics and photon statistics in bimodal quantum-dot micropillar lasers

Emission characteristics of quantum-dot micropillar lasers (QDMLs) are located at the intersection of nanophotonics and nonlinear dynamics, which provides an ideal platform for studying the optical interface between classical and quantum systems. In this work, a noise-induced bimodal QDML with orthogonal dual-mode outputs is modeled, and nonlinear dynamics, stochastic mode jumping and quantum statistics with the variation of stochastic noise intensity are investigated. Noise-induced effects lead to the emergence of two intensity bifurcation points for the strong and the weak mode, and the maximum output power of the strong mode becomes larger as the noise intensity increases. The anti-correlation of the two modes reaches the maximum at the second intensity bifurcation point. The dual-mode stochastic jumping frequency and effective bandwidth can exceed 100 GHz and 30 GHz under the noise-induced effect. Moreover, the noise-induced photon correlations of both modes simultaneously exhibit super-thermal bunching effects ($g^{(2)}(0)>2$) in the low injection current region. The $g^{(2)}(0)$-value of the strong mode can reach over 6 in the high injection current region. Photon bunching ($g^{(2)}(0)>1$) of both modes is observed over a wide range of noise intensities and injection currents. In the presence of the noise-induced effect, the photon number distribution of the strong or the weak mode is a mixture of Bose-Einstein and Poisson distributions. As the noise intensity increases, the photon number distribution of the strong mode is dominated by the Bose-Einstein distribution, and the proportion of the Poisson distribution is increased in the high injection current region, while that of the weak mode is reduced. Our results contribute to the development preparation of super-bunching quantum integrated light sources for improving the spatiotemporal resolution of quantum sensing measurements.

physics.optics↗

arXiv2023

Towards systematic intraday news screening: a liquidity-focused approach

News can convey bearish or bullish views on financial assets. Institutional investors need to evaluate automatically the implied news sentiment based on textual data. Given the huge amount of news articles published each day, most of which are neutral, we present a systematic news screening method to identify the ``true'' impactful ones, aiming for more effective development of news sentiment learning methods. Based on several liquidity-driven variables, including volatility, turnover, bid-ask spread, and book size, we associate each 5-min time bin to one of two specific liquidity modes. One represents the ``calm'' state at which the market stays for most of the time and the other, featured with relatively higher levels of volatility and trading volume, describes the regime driven by some exogenous events. Then we focus on the moments where the liquidity mode switches from the former to the latter and consider the news articles published nearby impactful. We apply naive Bayes on these filtered samples for news sentiment classification as an illustrative example. We show that the screened dataset leads to more effective feature capturing and thus superior performance on short-term asset return prediction compared to the original dataset.

q-fin.TR↗

arXiv2022

Multi-asset market making under the quadratic rough Heston

Given the promising results on joint modeling of SPX/VIX smiles of the recently introduced quadratic rough Heston model, we consider a multi-asset market making problem on SPX and its derivatives, e.g. VIX futures, SPX and VIX options. The market maker tries to maximize its profit from spread capturing while controlling the portfolio's inventory risk, which can be fully explained by the value change of SPX under the particular setting of the quadratic rough Heston model. The high dimensionality of the resulting optimization problem is relaxed by several approximations. An asymptotic closed-form solution can be obtained. The accuracy and relevance of the approximations are illustrated through numerical experiments.

q-fin.MF↗

arXiv2022

Improving Variational Autoencoders with Density Gap-based Regularization

Variational autoencoders (VAEs) are one of the powerful unsupervised learning frameworks in NLP for latent representation learning and latent-directed generation. The classic optimization goal of VAEs is to maximize the Evidence Lower Bound (ELBo), which consists of a conditional likelihood for generation and a negative Kullback-Leibler (KL) divergence for regularization. In practice, optimizing ELBo often leads the posterior distribution of all samples converge to the same degenerated local optimum, namely posterior collapse or KL vanishing. There are effective ways proposed to prevent posterior collapse in VAEs, but we observe that they in essence make trade-offs between posterior collapse and hole problem, i.e., mismatch between the aggregated posterior distribution and the prior distribution. To this end, we introduce new training objectives to tackle both two problems through a novel regularization based on the probabilistic density gap between the aggregated posterior distribution and the prior distribution. Through experiments on language modeling, latent space visualization and interpolation, we show that our proposed method can solve both problems effectively and thus outperforms the existing methods in latent-directed generation. To the best of our knowledge, we are the first to jointly solve the hole problem and the posterior collapse.

cs.LG↗