arXiv subjects

Jiarui Sun

Publications and source records attributed to Jiarui Sun.

At least 19 recordsLinked to original sources

arXiv2026

GradFlex: a Gradient-Based Method for Weak Lensing Flexion Measurement

Flexion, the second-order gravitational lensing distortion, probes small-scale mass distribution and complement the cosmic shear measurement. However, its detection is highly challenging due to the faint signal and the contamination arising from the intrinsic galaxy morphology. In this Letter, we present GradFlex, a new method that extracts flexion from the gradient field of galaxy surface brightness distributions without assuming a specific point-spread function model or galaxy morphology. Using surface-brightness conservation, we relate the pre- and post-lensing gradient fields and derive flexion estimators from ensemble-averaged statistics of the corresponding spins. Tests on numerical simulations across diverse galaxy morphologies and multiple noise realizations demonstrate its reliability under realistic observing conditions. Combined with shear measurements, it could tighten constraints on cluster density profiles and enable more detailed studies of substructure than shear alone.

astro-ph.CO↗

arXiv2026

TimeEvo: Failure-Driven Self-Evolution of a Time Series Agent

Time series agents answer analytical questions by calling external tools, and which tools they carry is decided by people before the agent runs. However, we identify two failures in this setup. Human-Agent Tool Misalignment: a library of 21 expert-curated tools helps on some tasks and hurts on others, dropping anomaly accuracy under every backbone we test. Silent Harm: one round of generic self-revision changes 147 answers and breaks 56 of them, while the final score moves by less than a point. Both follow from the same gap: whether a tool helps is decided question by question at runtime, while tools are supplied in advance and judged by a single average. To address this, we propose TimeEvo, which clusters an agent's diagnosed failures into capability gaps, plans a measurement for each, synthesizes evidence-only tools that fill them, and admits the candidate library only through a paired admission gate. Experiments on ten time series QA tasks and three backbones show that TimeEvo, starting from an empty library, improves accuracy on every task and every backbone, and that a library grown on a cheap model still gains when it is installed into stronger ones. Code is available at https://github.com/Muyiiiii/TimeEvo.

cs.AI↗

arXiv2026

LastOPD: Taming Collapse in Latent On-Policy Distillation

On-policy distillation (OPD) corrects a student on the responses it writes, but its signal is the teacher's next-token distribution: it tells the student what the teacher says but misses how it thinks. Latent supervision promises the missing part by aligning the student's latent states to the teacher's. Recent methods such as OPRD bring this signal into on-policy distillation. However, we observe two failures of this recipe when distilling Qwen3-4B and Qwen3-8B into Qwen3-1.7B-Base. Early gain, late collapse: latent supervision alone lifts MATH-500 accuracy from 25 to 46 in 10 steps, but subsequent training degrades performance down to 11 with no recovery. Better alignment, worse behavior: although the alignment metric steadily improves throughout this collapse, the most aligned model turns out to be the worst performing. Further analysis suggests a mismatch in how the latent signal is applied: layers paired by depth play different roles in the two models, so continued alignment may pull the student toward teacher states it cannot understand. To address this, we propose LastOPD, which applies the latent signal only at the last-layer state, the common interface both LM heads read, and only during a 10-step crossfade into token-level OPD. This keeps the useful part of the latent signal and hands the student to token-level supervision before the collapse sets in. Extensive experiments show that LastOPD improves MATH-500 over token-only OPD by 5.55 and 4.02 points with the 4B and 8B teachers, leads on most held-out datasets, and reaches the final score of token-only OPD in about half the steps. Code is available at https://github.com/Muyiiiii/LastOPD.

cs.LG↗

arXiv2026

CLARity: Reasoning Consistency Alone Can Teach Reinforced Experts

Training expert LLMs in domains with scarce data is difficult, often relying on multiple-choice questions (MCQs). However, standard outcome-based reinforcement learning (RL) on MCQs is risky. While it may improve accuracy, we observe it often degrades reasoning quality such as logical consistency. Existing solutions to supervise reasoning, such as large-scale Process Reward Models (PRMs), are prohibitively expensive. To address this, we propose CLARity, a cost-effective RL framework that enhances reasoning quality using only a small, general-purpose LLM. CLARity integrates a consistency-aware reward mechanism with a 2-stage refine-then-monitor training pipeline to enhance reasoning consistency, and a dynamic data reformulation strategy to to better exploit limited data. Experiments demonstrate that CLARity improves response consistency by 16.5% and accuracy by 7.5% over baselines. Human evaluations further confirm holistic improvements in coherence and professionalism. Thus, CLARity offers a generalizable solution that enables smaller models to effectively guide expert models by reasoning consistency. Our code is open sourced at: https://github.com/Infinite-set/CLARity

cs.CL↗

arXiv2026

Calibrating Self-Weighted Shear Estimation via Field Distortion

Accurate cosmic shear measurement is the key to fully realize the scientific potential of large scale galaxy surveys. The Fourier\_Quad shear measurement method has been significantly developed to avoid biases caused by various factors, including the point spread function (PSF), photon noise, pixelation effect, selection effects, etc.. Shear statistics has also been optimized by using the PDF-SYM method (symmetrization of the Probability Distribution Function of the shear estimators) to achieve the minimal statistical error, without introducing systematic error. Nevertheless, the iterative nature of the PDF-SYM method makes it computationally expensive for massive datasets. To substantially improve efficiency, we introduce the Self-Weighted Shear Estimation (SWSE) method, which employs a specific self-weighting scheme to naturally suppress the shape noise as well as to balance the contributions from the bright and faint galaxies. Because SWSE is intrinsically biased, the primary focus of this work is to robustly evaluate whether the field-distortion test (FD test, a way to calibrate the shear bias on real data) can accurately recover and correct these inherent multiplicative and additive biases. Using both simulated and real survey data, we demonstrate that the FD test can successfully calibrate SWSE on galaxy-galaxy lensing measurements, enabling it to achieve statistical precision comparable to PDF-SYM. Our results establish the combination of FD test and SWSE as a robust shear estimation approach for forthcoming large-scale weak lensing surveys.

astro-ph.CO↗

arXiv2026

TREASURE: The Visa Payment Foundation Model for High-Volume Transaction Understanding

Payment networks form the backbone of modern commerce, generating high volumes of transaction records from daily activities. Properly modeling this data can enable applications such as abnormal behavior detection and consumer-level insights for hyper-personalized experiences, ultimately improving people's lives. In this paper, we present TREASURE, TRansformer Engine As Scalable Universal transaction Representation Encoder, a multipurpose transformer-based foundation model specifically designed for transaction data. The model simultaneously captures both consumer behavior and payment network signals (such as response codes and system flags), providing comprehensive information necessary for applications like accurate recommendation systems and abnormal behavior detection. Verified with industry-grade datasets, TREASURE features three key capabilities: 1) an input module with dedicated sub-modules for static and dynamic attributes, enabling more efficient training and inference; 2) an efficient and effective training paradigm for predicting high-cardinality categorical attributes; and 3) demonstrated effectiveness as both a standalone model that increases abnormal behavior detection performance by 111% over production systems and an embedding provider that enhances recommendation models by 104%. We present key insights from extensive ablation studies, benchmarks against production models, and case studies, highlighting valuable knowledge gained from developing TREASURE.

cs.LG↗

arXiv2026

S2Act: Simple Spiking Actor

Spiking neural networks (SNNs) and biologically-inspired learning mechanisms are attractive in mobile robotics, where the size and performance of onboard neural network policies are constrained by power and computational budgets. Existing SNN approaches, such as population coding, reward modulation, and hybrid artificial neural network (ANN)-SNN architectures, have shown promising results; however, they face challenges in complex, highly stochastic environments due to SNN sensitivity to hyperparameters and inconsistent gradient signals. To address these challenges, we propose simple spiking actor (S2Act), a computationally lightweight framework that deploys an RL policy using an SNN in three steps: (1) architect an actor-critic model based on an approximated network of rate-based spiking neurons, (2) train the network with gradients using compatible activation functions, and (3) transfer the trained weights into physical parameters of rate-based leaky integrate-and-fire (LIF) neurons for inference and deployment. By globally shaping LIF neuron parameters such that their rate-based responses approximate ReLU activations, S2Act effectively mitigates the vanishing gradient problem, while pre-constraining LIF response curves reduces reliance on complex SNN-specific hyperparameter tuning. We demonstrate our method in two multi-agent stochastic environments (capture-the-flag and parking) that capture the complexity of multi-robot interactions, and deploy our trained policies on physical TurtleBot platforms using Intel's Loihi neuromorphic hardware. Our experimental results show that S2Act outperforms relevant baselines in task performance and real-time inference in nearly all considered scenarios, highlighting its potential for rapid prototyping and efficient real-world deployment of SNN-based RL policies.

cs.MA↗

arXiv2026

TransactionGPT

We present TransactionGPT (TGPT), a foundation model for consumer transaction data within one of the world's largest payment networks. TGPT is designed to understand and generate transaction trajectories while simultaneously supporting a variety of downstream prediction and classification tasks. We introduce a novel 3D-Transformer architecture specifically tailored for capturing the complex dynamics in payment transaction data. This architecture incorporates design innovations that enhance modality fusion and computational efficiency, while seamlessly enabling joint optimization with downstream objectives. Trained on billion-scale real-world transactions, TGPT significantly improves downstream anomaly transaction detection performance against a competitive production model and exhibits advantages over baselines in generating future transactions. We conduct extensive empirical evaluations utilizing a diverse collection of company transaction datasets spanning multiple downstream tasks, thereby enabling a thorough assessment of TGPT's effectiveness and efficiency in comparison to established methodologies. Furthermore, we examine the incorporation of LLM-derived embeddings within TGPT and benchmark its performance against fine-tuned LLMs, demonstrating that TGPT achieves superior predictive accuracy as well as faster training and inference. We anticipate that the architectural innovations and practical guidelines from this work will advance foundation models for transaction-like data and catalyze future research in this emerging field.

cs.LG↗

arXiv2026

Understanding LLM Evaluator Behavior: A Structured Multi-Evaluator Framework for Merchant Risk Assessment

Large Language Models (LLMs) are increasingly used as evaluators of reasoning quality, yet their reliability and bias in payments-risk settings remain poorly understood. We introduce a structured multi-evaluator framework for assessing LLM reasoning in Merchant Category Code (MCC)-based merchant risk assessment, combining a five-criterion rubric with Monte-Carlo scoring to evaluate rationale quality and evaluator stability. Five frontier LLMs generate and cross-evaluate MCC risk rationales under attributed and anonymized conditions. To establish a judge-independent reference, we introduce a consensus-deviation metric that eliminates circularity by comparing each judge's score to the mean of all other judges, yielding a theoretically grounded measure of self-evaluation and cross-model deviation. Results reveal substantial heterogeneity: GPT-5.1 and Claude 4.5 Sonnet show negative self-evaluation bias (-0.33, -0.31), while Gemini-2.5 Pro and Grok 4 display positive bias (+0.77, +0.71), with bias attenuating by 25.8 percent under anonymization. Evaluation by 26 payment-industry experts shows LLM judges assign scores averaging +0.46 points above human consensus, and that the negative bias of GPT-5.1 and Claude 4.5 Sonnet reflects closer alignment with human judgment. Ground-truth validation using payment-network data shows four models exhibit statistically significant alignment (Spearman rho = 0.56 to 0.77), confirming that the framework captures genuine quality. Overall, the framework provides a replicable basis for evaluating LLM-as-a-judge systems in payment-risk workflows and highlights the need for bias-aware protocols in operational financial settings.

cs.AI↗

arXiv2026

MiMo-V2-Flash Technical Report

We present MiMo-V2-Flash, a Mixture-of-Experts (MoE) model with 309B total parameters and 15B active parameters, designed for fast, strong reasoning and agentic capabilities. MiMo-V2-Flash adopts a hybrid attention architecture that interleaves Sliding Window Attention (SWA) with global attention, with a 128-token sliding window under a 5:1 hybrid ratio. The model is pre-trained on 27 trillion tokens with Multi-Token Prediction (MTP), employing a native 32k context length and subsequently extended to 256k. To efficiently scale post-training compute, MiMo-V2-Flash introduces a novel Multi-Teacher On-Policy Distillation (MOPD) paradigm. In this framework, domain-specialized teachers (e.g., trained via large-scale reinforcement learning) provide dense and token-level reward, enabling the student model to perfectly master teacher expertise. MiMo-V2-Flash rivals top-tier open-weight models such as DeepSeek-V3.2 and Kimi-K2, despite using only 1/2 and 1/3 of their total parameters, respectively. During inference, by repurposing MTP as a draft model for speculative decoding, MiMo-V2-Flash achieves up to 3.6 acceptance length and 2.6x decoding speedup with three MTP layers. We open-source both the model weights and the three-layer MTP weights to foster open research and community collaboration.

cs.CL↗

arXiv2025

A Unified Inference Method for FROC-type Curves and Related Summary Indices

Free-response observer performance studies are of great importance for accuracy evaluation and comparison in tasks related to the detection and localization of multiple targets or signals. The free-response receiver operating characteristic (FROC) curve and many similar curves based on the free-response observer performance assessment data are important tools to display the accuracy of detection under different thresholds. The true positive rate at a fixed false positive rate and summary indices such as the area under the FROC curve are also commonly used as the figures of merit in the statistical evaluation of these studies. Motivated by a free-response observer performance assessment research of a Software as a Medical Device (SaMD), we propose a unified method based on the initial-detection-and-candidate model to simultaneously estimate a smooth curve and derive confidence intervals for summary indices and the true positive rate at a fixed false positive rate. A maximum likelihood estimator is proposed and its asymptotic normality property is derived. Confidence intervals are constructed based on the asymptotic normality of our maximum likelihood estimator. Simulation studies are conducted to evaluate the finite sample performance of the proposed method. We apply the proposed method to evaluate the diagnostic performance of the SaMD for detecting pulmonary lesions.

stat.ME↗

arXiv2025

TiCT: A Synthetically Pre-Trained Foundation Model for Time Series Classification

The ubiquity of time series data creates a strong demand for general-purpose foundation models, yet developing them for classification remains a significant challenge, largely due to the high cost of labeled data. Foundation models capable of in-context learning (ICL) offer a powerful solution, adapting to new tasks with minimal examples and reducing the need for extensive retraining. However, prior work on large-scale time series models has predominantly focused on forecasting, leaving a critical gap for versatile, fine-tuning-free classification. To address this, we introduce TiCT (Time-series in-Context Transformer), a transformer-based model pre-trained exclusively on synthetic data to perform in-context classification. We make two primary technical contributions: 1) a novel architecture featuring a scalable bit-based label encoding and a special output attention mechanism to handle an arbitrary number of classes; and 2) a synthetic pre-training framework that combines a Mixup-inspired process with data augmentation to foster generalization and noise invariance. Extensive evaluations on the UCR Archive show that TiCT achieves competitive performance against state-of-the-art supervised methods. Crucially, this is accomplished using only in-context examples at inference time, without updating a single model weight.

cs.LG↗

arXiv2025

Precision Measurement of Large Shear Signals

So far, estimators of galaxy shape distortions are only carefully studied perturbatively in the case of small shear signals, mainly for weak lensing science. However, in the neighborhood of massive foreground clusters, a large number of background galaxies can be significantly distorted. The measurement of such large shear signals could be quite nontrivial under general observing conditions, i.e., in the presence of the point spread function (PSF) and noise. In this work, we propose a non-perturbative method to exactly recover large shear signals ($\gtrsim 0.5$) under general conditions. We test the method on simulated galaxy images, and find that it is accurate down to the very faint end. This new method is particularly useful for more accurate recovery of the shear distribution in the neighborhood of massive foreground clusters, thereby improving the modeling of the underlying dark matter halo properties.

astro-ph.CO↗

arXiv2025

Towards Efficient Large Scale Spatial-Temporal Time Series Forecasting via Improved Inverted Transformers

Time series forecasting at scale presents significant challenges for modern prediction systems, particularly when dealing with large sets of synchronized series, such as in a global payment network. In such systems, three key challenges must be overcome for accurate and scalable predictions: 1) emergence of new entities, 2) disappearance of existing entities, and 3) the large number of entities present in the data. The recently proposed Inverted Transformer (iTransformer) architecture has shown promising results by effectively handling variable entities. However, its practical application in large-scale settings is limited by quadratic time and space complexity ($O(N^2)$) with respect to the number of entities $N$. In this paper, we introduce EiFormer, an improved inverted transformer architecture that maintains the adaptive capabilities of iTransformer while reducing computational complexity to linear scale ($O(N)$). Our key innovation lies in restructuring the attention mechanism to eliminate redundant computations without sacrificing model expressiveness. Additionally, we incorporate a random projection mechanism that not only enhances efficiency but also improves prediction accuracy through better feature representation. Extensive experiments on the public LargeST benchmark dataset and a proprietary large-scale time series dataset demonstrate that EiFormer significantly outperforms existing methods in both computational efficiency and forecasting accuracy. Our approach enables practical deployment of transformer-based forecasting in industrial applications where handling time series at scale is essential.

cs.LG↗

arXiv2025

A Comprehensive Catalog of Emission-line Nebulae, Star Clusters, and Supergiants in M31 from the LAMOST Spectroscopic Survey

Spectroscopic observations of various tracers in nearby galaxies, such as Andromeda (M31), play a crucial role in identifying and classifying individual stellar populations and nebular objects, thereby enhancing our understanding of galactic composition, environment, and dynamics as well as stellar evolution. While the LAMOST (Large Sky Area Multi-Object Fibre Spectroscopic Telescope) survey of M31 has produced extensive datasets, a comprehensive catalog of emission-line nebulae, star clusters, and supergiants is yet to be completed. In this paper, we present a final catalog of 384 emission-line nebulae, 380 star clusters, and 375 supergiants and candidates in M31, as carefully selected and identified from the LAMOST spectroscopic database. These objects were classified using a random forest algorithm, followed by thorough visual examinations of their spectral characteristics as well as morphologies revealed by archive images. For emission-line nebulae, we measured radial velocities and relative fluxes of emission lines, enabling further classification of planetary nebulae and HII regions. Additionally, we identified 245 emission-line nebulae in M33. This work lays the data foundation for the study of M31, and offers valuable tracers to investigate M31's structure and evolution.

astro-ph.GA↗

arXiv2024

MOOSS: Mask-Enhanced Temporal Contrastive Learning for Smooth State Evolution in Visual Reinforcement Learning

In visual Reinforcement Learning (RL), learning from pixel-based observations poses significant challenges on sample efficiency, primarily due to the complexity of extracting informative state representations from high-dimensional data. Previous methods such as contrastive-based approaches have made strides in improving sample efficiency but fall short in modeling the nuanced evolution of states. To address this, we introduce MOOSS, a novel framework that leverages a temporal contrastive objective with the help of graph-based spatial-temporal masking to explicitly model state evolution in visual RL. Specifically, we propose a self-supervised dual-component strategy that integrates (1) a graph construction of pixel-based observations for spatial-temporal masking, coupled with (2) a multi-level contrastive learning mechanism that enriches state representations by emphasizing temporal continuity and change of states. MOOSS advances the understanding of state dynamics by disrupting and learning from spatial-temporal correlations, which facilitates policy learning. Our comprehensive evaluation on multiple continuous and discrete control benchmarks shows that MOOSS outperforms previous state-of-the-art visual RL methods in terms of sample efficiency, demonstrating the effectiveness of our method. Our code is released at https://github.com/jsun57/MOOSS.

cs.CV↗

arXiv2024

CoMusion: Towards Consistent Stochastic Human Motion Prediction via Motion Diffusion

Stochastic Human Motion Prediction (HMP) aims to predict multiple possible future human pose sequences from observed ones. Most prior works learn motion distributions through encoding-decoding in the latent space, which does not preserve motion's spatial-temporal structure. While effective, these methods often require complex, multi-stage training and yield predictions that are inconsistent with the provided history and can be physically unrealistic. To address these issues, we propose CoMusion, a single-stage, end-to-end diffusion-based stochastic HMP framework. CoMusion is inspired from the insight that a smooth future pose initialization improves prediction performance, a strategy not previously utilized in stochastic models but evidenced in deterministic works. To generate such initialization, CoMusion's motion predictor starts with a Transformer-based network for initial reconstruction of corrupted motion. Then, a graph convolutional network (GCN) is employed to refine the prediction considering past observations in the discrete cosine transformation (DCT) space. Our method, facilitated by the Transformer-GCN module design and a proposed variance scheduler, excels in predicting accurate, realistic, and consistent motions, while maintaining appropriate diversity. Experimental results on benchmark datasets demonstrate that CoMusion surpasses prior methods across metrics, while demonstrating superior generation quality. Our Code is released at https://github.com/jsun57/CoMusion/ .

cs.CV↗

arXiv2024

Revealing the Power of Masked Autoencoders in Traffic Forecasting

Traffic forecasting, crucial for urban planning, requires accurate predictions of spatial-temporal traffic patterns across urban areas. Existing research mainly focuses on designing complex models that capture spatial-temporal dependencies among variables explicitly. However, this field faces challenges related to data scarcity and model stability, which results in limited performance improvement. To address these issues, we propose Spatial-Temporal Masked AutoEncoders (STMAE), a plug-and-play framework designed to enhance existing spatial-temporal models on traffic prediction. STMAE consists of two learning stages. In the pretraining stage, an encoder processes partially visible traffic data produced by a dual-masking strategy, including biased random walk-based spatial masking and patch-based temporal masking. Subsequently, two decoders aim to reconstruct the masked counterparts from both spatial and temporal perspectives. The fine-tuning stage retains the pretrained encoder and integrates it with decoders from existing backbones to improve forecasting accuracy. Our results on traffic benchmarks show that STMAE can largely enhance the forecasting capabilities of various spatial-temporal models.

cs.LG↗