arXiv subjects

Ke Hu

Publications and source records attributed to Ke Hu.

At least 19 recordsLinked to original sources

arXiv2026

From Language to Behavior: Scaling Sequence Transformers for Industrial Recommendation Ranking with Rec-Native Designs

Scaling Transformers has driven large gains in language modeling, but transplanting this to behavior-sequence modeling in production ranking is challenging: recommendation differs in signal quality, where behavior sequences are noisy, temporally irregular, and sparsely supervised, and in computation asymmetry, where each request scores many candidates against one shared user history under tight latency budgets. We propose ReST, a recommendation-native Transformer scaling framework. For signal quality, it introduces a sequence encoder with dual-gated attention, rotary positional and temporal embedding, stabilized residual normalization, and training-only auxiliary objectives. For computation asymmetry, it factorizes ranking into a heavy reusable encoder and a lightweight cross decoder with projection-free KV attention and token-specific parameterization, coupling user-level shared-prefix training with shared-prefix serving for compute-once, decode-many-times ranking. Across industrial and public benchmarks, ReST achieves higher accuracy and scales more consistently along sequence length, depth, and width, where LLM-style Transformer blocks saturate. A one-week online A/B test on a production advertising platform improves online AUC by 1.31% and lifts a core revenue metric by 11.93% within a 50 ms P99 budget; ReST has since been fully deployed in production, showing that behavior-sequence scaling remains a promising, under-exploited axis for production ranking.

cs.IR↗

arXiv2026

SN1987A constraints on the neutrino-dark-fermion interaction from resonant scattering with C$\nu$B

Neutrino self-interactions mediated by a light scalar offer a compelling resolution to cosmological tensions and may naturally arise in neutrino mass generation mechanisms. When the scalar also couples to a light dark-sector fermion, supernova neutrinos can resonantly annihilate with the cosmic neutrino background (C$\nu$B) into invisible dark radiation, depleting the flux en route to Earth. We use this depletion to constrain the neutrino-scalar coupling from the SN1987A data. Within a Bayesian framework we analyze the data for two supernova neutrino emission models, a parameterized model and a 2D hydrodynamic simulation, and for two coupling types, a mass-independent one and a mass-proportional one. Our results show that new-physics limits from SN1987A cannot be quoted independently of the heavy-flavor emission, the flavor conversion, or the coupling structure. Finally, we forecast that a future high-statistics burst recorded by Hyper-Kamiokande would restore a meaningful upper bound even when flavor conversion is included.

hep-ph↗

arXiv2026

VoiceChat-TTS: A Low-Latency Continuous Speech Synthesis Model for Interactive Agents

Spoken dialogue is a natural form of human--computer interaction, yet most speech language models remain limited to turn-based operation and lack real-time adaptability, such as user barge-in. Recent duplex speech-to-speech and speech-to-text models reduce latency by replacing multi-stage pipelines, but often compromise speech quality because accurate ASR, interruption handling, and high-fidelity synthesis must be optimized jointly. We propose VoiceChat-TTS, a low-latency, continuous, and streamable text-to-speech model for interactive agents. VoiceChat-TTS is driven directly by LLM text-token streams, supports explicit interruption via control tokens, and produces silence when no textual input is available. The model enables always-on, responsive speech generation while preserving modularity and high speech quality, and it supports mid-utterance interruptions without resetting the KV cache.

eess.AS↗

arXiv2026

Steering Generative Reinforcement Learning into Stable Robotic Controller

Diffusion and flow-based generative policies provide a powerful policy class for reinforcement learning by inducing rich stochastic exploration through iterative action generation. However, the stochasticity of diffusion policies is not suitable for stable and precise control in high-dimensional robotic systems, where small action variations can accumulate into inconsistent motion and reduced robustness. To address this issue, we propose SteerGenPO, a latent-space reinforcement learning framework that steers a trained generative policy into a robust deterministic robotic controller. The key idea is to replace stochastic latent sampling of the trained generative policy with a learned latent actor that predicts a state-dependent latent input for the generative policies. This separates exploration and control: stochastic generative sampling provides diverse action proposals during policy learning, while deterministic latent steering provides stable and adaptive control at deployment. We evaluate SteerGenPO on six Isaac Lab benchmarks and a Unitree G1 locomotion task. The results show SteerGenPO improves over both classical RL and generative RL baselines, while its deterministic latent steering produces more stable inference-time behaviors and more reliable command responses.

cs.RO↗

arXiv2026

GenPO++: Generative Policy Optimization with Jacobian-free Likelihood Ratios

Generative policies provide expressive and multimodal action distributions, making them attractive for reinforcement learning (RL) in complex continuous-control tasks. Among them, flow-based policies are especially appealing because they generate actions through deterministic transport maps. However, applying such generative policies to likelihood-based on-policy learning remains limited by the difficulty of evaluating the probability of executed actions. Existing flow RL methods either replace the true action-density ratio with approximate surrogates, which can introduce biased updates, or recover exact likelihoods through dummy-action augmentation, which enlarges the policy space and increases computation. In this work, we propose GenPO++, a reversible generative policy optimization framework that uses history states as auxiliary memory in a high-order reversible ODE solver, yielding exact inversion without changing the original action dimension. The resulting generative policy map has a log-determinant determined only by fixed solver coefficients, enabling exact and Jacobian-free likelihood-ratio computation. This design preserves the expressiveness of generative flow policies while avoiding both action ratio bias and dummy-action overhead. We evaluate GenPO++ on large-scale simulated control, fine-tuning, and real-world robotic manipulation tasks, where it achieves competitive or superior performance over state-of-the-art on-policy RL methods, while improving training stability and computational efficiency.

cs.LG↗

arXiv2026

Sample-Efficient Diffusion-based Reinforcement Learning with Critic Guidance

Recent advances in reinforcement learning (RL) have achieved great successes by leveraging the multimodality and exploration capability of diffusion policies. Among these approaches, one representative branch focuses on the sampling-based policy optimization. This design enables better exploration capability of the diffusion model, particularly at the beginning of training, but suffer from low exploitation in Q-value information, resulting in a slow policy convergence. Another branch pays attention to gradient-based policy optimization, which sufficiently exploits the gradient of the Q function yet tends to collapse into a unimodal policy with low diversity. To address this issue, we propose CGPO, \textbf{C}ritic-\textbf{G}uided diffusion \textbf{P}olicy \textbf{O}ptimization, which effectively balances exploration and exploitation with the training-free guidance technique integrated into the denoising process of diffusion policy. Concretely, CGPO steers action generation toward high-value regions defined by the critic network and uses the guided actions as regression objectives. In this manner, CGPO reduces the time required to obtain high-quality actions and improves final performance with better balance between the exploration-exploitation tradeoff. We validate the effectiveness of CGPO on 5 MuJoCo locomotion tasks, and CGPO achieves state-of-the-art performance compared with existing diffusion-based RL methods. Notably, CGPO is the first success to incorporate diffusion policy into real-world RL, with its superior performance on Franka robot arm grasping tasks. Our official page is released at https://dingsht.tech/cgpo-webpage.

cs.RO↗

arXiv2026

Explainable AML Triage with LLMs: Evidence Retrieval and Counterfactual Checks

Anti-money laundering (AML) transaction monitoring generates large volumes of alerts that must be rapidly triaged by investigators under strict audit and governance constraints. While large language models (LLMs) can summarize heterogeneous evidence and draft rationales, unconstrained generation is risky in regulated workflows due to hallucinations, weak provenance, and explanations that are not faithful to the underlying decision. We propose an explainable AML triage framework that treats triage as an evidence-constrained decision process. Our method combines (i) retrieval-augmented evidence bundling from policy/typology guidance, customer context, alert triggers, and transaction subgraphs, (ii) a structured LLM output contract that requires explicit citations and separates supporting from contradicting or missing evidence, and (iii) counterfactual checks that validate whether minimal, plausible perturbations lead to coherent changes in both the triage recommendation and its rationale. We evaluate on public synthetic AML benchmarks and simulators and compare against rules, tabular and graph machine-learning baselines, and LLM-only/RAG-only variants. Results show that evidence grounding substantially improves auditability and reduces numerical and policy hallucination errors, while counterfactual validation further increases decision-linked explainability and robustness, yielding the best overall triage performance (PR-AUC 0.75; Escalate F1 0.62) and strong provenance and faithfulness metrics (citation validity 0.98; evidence support 0.88; counterfactual faithfulness 0.76). These findings indicate that governed, verifiable LLM systems can provide practical decision support for AML triage without sacrificing compliance requirements for traceability and defensibility.

cs.AI↗

arXiv2026

Distributional Reinforcement Learning with Diffusion Bridge Critics

Recent advances in diffusion-based reinforcement learning (RL) methods have demonstrated promising results in a wide range of continuous control tasks. However, existing works in this field focus on the application of diffusion policies while leaving the diffusion critics unexplored. In fact, since policy optimization fundamentally relies on the critic, accurate value estimation is far more important than policy expressiveness. Furthermore, given the stochasticity of most reinforcement learning tasks, it has been confirmed that the critic is more appropriately depicted with a distributional model. Motivated by these points, we propose a novel distributional RL method with Diffusion Bridge Critics (DBC). DBC directly models the inverse cumulative distribution function (CDF) of the Q value. This allows us to accurately capture the value distribution and prevents it from collapsing into a trivial Gaussian distribution owing to the strong distribution-matching capability of the diffusion bridge. Moreover, we further derive an analytic integral formula to address discretization errors in DBC, which is essential in value estimation. To our knowledge, DBC is the first work to employ the diffusion bridge model as the critic. Notably, DBC is also a plug-and-play component and can be integrated into most existing RL frameworks. Experimental results on MuJoCo robot control benchmarks demonstrate the superiority of DBC compared with previous distributional critic models.

cs.LG↗

arXiv2025

Deep learning-driven atmospheric parameter prediction for hot subdwarf stars with synthetic and observed spectra

We design a convolutional neural network (CNN) incorporating channel attention and spatial attention mechanisms to predict atmospheric parameters of hot subdwarfs. The experimental dataset comprises spectra at nine distinct signal-to-noise ratio (SNR) levels, with each SNR level containing 11 396 synthetic spectra and 945 observed spectra. The trained deep learning models achieves mean absolute errors (AME) in predicting hot subdwarf atmospheric parameters of 730 K for effective temperature (Teff ), 0.09 dex for surface gravity (log g), and 0.03 dex for helium abundance (log(nHe/nH)), respectively, which reaches the accuracy of traditional spectral fitting methods. Utilizing the trained deep learning models and low-resolution spectra from LAMOST DR12, we confirm 1512 hot subdwarfs from the catalog of hot subdwarf candidates, of which 291 are newly identified. Our results demonstrate that the deep learning model not only achieves accuracy comparable to traditional methods in obtaining hot subdwarf atmospheric parameters, but also far exceeds them in speed and efficiency, making it particularly suitable for the analysis of large datasets of hot subdwarf spectra.

astro-ph.SR↗

arXiv2025

Defects and Inconsistencies in Solar Flare Data Sources: Implications for Machine Learning Forecasting

Machine learning models for forecasting solar flares have been trained and evaluated using a variety of data sources, including Space Weather Prediction Center (SWPC) operational and science-quality data. Typically, data from these sources is minimally processed before being used to train and validate a forecasting model. However, predictive performance can be affected if defects and inconsistencies between these data sources are ignored. For a set of commonly used data sources, along with the software that queries and outputs processed data, we identify their defects and inconsistencies, quantify their extent, and show how they can affect predictions from data-driven machine-learning forecasting models. We also outline procedures for fixing these issues or at least mitigating their impacts. Finally, based on thorough comparisons of the effects of data sources on the trained forecasting model's predictive skill scores, we offer recommendations for using different data products in operational forecasting.

astro-ph.SR↗

arXiv2025

FedTopo: Topology-Informed Representation Alignment in Federated Learning under Non-I.I.D. Conditions

Current federated-learning models deteriorate under heterogeneous (non-I.I.D.) client data, as their feature representations diverge and pixel- or patch-level objectives fail to capture the global topology which is essential for high-dimensional visual tasks. We propose FedTopo, a framework that integrates Topological-Guided Block Screening (TGBS) and Topological Embedding (TE) to leverage topological information, yielding coherently aligned cross-client representations by Topological Alignment Loss (TAL). First, Topology-Guided Block Screening (TGBS) automatically selects the most topology-informative block, i.e., the one with maximal topological separability, whose persistence-based signatures best distinguish within- versus between-class pairs, ensuring that subsequent analysis focuses on topology-rich features. Next, this block yields a compact Topological Embedding, which quantifies the topological information for each client. Finally, a Topological Alignment Loss (TAL) guides clients to maintain topological consistency with the global model during optimization, reducing representation drift across rounds. Experiments on Fashion-MNIST, CIFAR-10, and CIFAR-100 under four non-I.I.D. partitions show that FedTopo accelerates convergence and improves accuracy over strong baselines.

cs.LG↗

arXiv2025

HOID-R1: Reinforcement Learning for Open-World Human-Object Interaction Detection Reasoning with Multimodal Large Language Model

Understanding and recognizing human-object interaction (HOI) is a pivotal application in AR/VR and robotics. Recent open-vocabulary HOI detection approaches depend exclusively on large language models for richer textual prompts, neglecting their inherent 3D spatial understanding capabilities. To address this shortcoming, we introduce HOID-R1, the first HOI detection framework that integrates chain-of-thought (CoT) guided supervised fine-tuning (SFT) with group relative policy optimization (GRPO) within a reinforcement learning (RL) paradigm. Specifically, we initially apply SFT to imbue the model with essential reasoning capabilities, forcing the model to articulate its thought process in the output. Subsequently, we integrate GRPO to leverage multi-reward signals for policy optimization, thereby enhancing alignment across diverse modalities. To mitigate hallucinations in the CoT reasoning, we introduce an "MLLM-as-a-judge" mechanism that supervises the CoT outputs, further improving generalization. Extensive experiments show that HOID-R1 achieves state-of-the-art performance on HOI detection benchmarks and outperforms existing methods in open-world generalization to novel scenarios.

cs.CV↗

arXiv2025

SpeechIQ: Speech-Agentic Intelligence Quotient Across Cognitive Levels in Voice Understanding by Large Language Models

We introduce Speech-based Intelligence Quotient (SIQ) as a new form of human cognition-inspired evaluation pipeline for voice understanding large language models, LLM Voice, designed to assess their voice understanding ability. Moving beyond popular voice understanding metrics such as word error rate (WER), SIQ examines LLM Voice across three cognitive levels motivated by Bloom's Taxonomy: (1) Remembering (i.e., WER for verbatim accuracy); (2) Understanding (i.e., similarity of LLM's interpretations); and (3) Application (i.e., QA accuracy for simulating downstream tasks). We demonstrate that SIQ not only quantifies voice understanding abilities but also provides unified comparisons between cascaded methods (e.g., ASR LLM) and end-to-end models, identifies annotation errors in existing benchmarks, and detects hallucinations in LLM Voice. Our framework represents a first-of-its-kind intelligence examination that bridges cognitive principles with voice-oriented benchmarks, while exposing overlooked challenges in multi-modal training. Our code and data will be open source to encourage future studies.

cs.CL↗

arXiv2025

GenPO: Generative Diffusion Models Meet On-Policy Reinforcement Learning

Recent advances in reinforcement learning (RL) have demonstrated the powerful exploration capabilities and multimodality of generative diffusion-based policies. While substantial progress has been made in offline RL and off-policy RL settings, integrating diffusion policies into on-policy frameworks like PPO remains underexplored. This gap is particularly significant given the widespread use of large-scale parallel GPU-accelerated simulators, such as IsaacLab, which are optimized for on-policy RL algorithms and enable rapid training of complex robotic tasks. A key challenge lies in computing state-action log-likelihoods under diffusion policies, which is straightforward for Gaussian policies but intractable for flow-based models due to irreversible forward-reverse processes and discretization errors (e.g., Euler-Maruyama approximations). To bridge this gap, we propose GenPO, a generative policy optimization framework that leverages exact diffusion inversion to construct invertible action mappings. GenPO introduces a novel doubled dummy action mechanism that enables invertibility via alternating updates, resolving log-likelihood computation barriers. Furthermore, we also use the action log-likelihood for unbiased entropy and KL divergence estimation, enabling KL-adaptive learning rates and entropy regularization in on-policy updates. Extensive experiments on eight IsaacLab benchmarks, including legged locomotion (Ant, Humanoid, Anymal-D, Unitree H1, Go2), dexterous manipulation (Shadow Hand), aerial control (Quadcopter), and robotic arm tasks (Franka), demonstrate GenPO's superiority over existing RL baselines. Notably, GenPO is the first method to successfully integrate diffusion policies into on-policy RL, unlocking their potential for large-scale parallelized training and real-world robotic deployment.

cs.LG↗

arXiv2025

Word Level Timestamp Generation for Automatic Speech Recognition and Translation

We introduce a data-driven approach for enabling word-level timestamp prediction in the Canary model. Accurate timestamp information is crucial for a variety of downstream tasks such as speech content retrieval and timed subtitles. While traditional hybrid systems and end-to-end (E2E) models may employ external modules for timestamp prediction, our approach eliminates the need for separate alignment mechanisms. By leveraging the NeMo Forced Aligner (NFA) as a teacher model, we generate word-level timestamps and train the Canary model to predict timestamps directly. We introduce a new <|timestamp|> token, enabling the Canary model to predict start and end timestamps for each word. Our method demonstrates precision and recall rates between 80% and 90%, with timestamp prediction errors ranging from 20 to 120 ms across four languages, with minimal WER degradation. Additionally, we extend our system to automatic speech translation (AST) tasks, achieving timestamp prediction errors around 200 milliseconds.

cs.CL↗

arXiv2025

SALM-Duplex: Efficient and Direct Duplex Modeling for Speech-to-Speech Language Model

Spoken dialogue is an intuitive form of human-computer interaction, yet current speech language models often remain constrained to turn-based exchanges, lacking real-time adaptability such as user barge-in. We propose a novel duplex speech to speech (S2S) architecture featuring continuous user inputs and codec agent outputs with channel fusion that directly models simultaneous user and agent streams. Using a pretrained streaming encoder for user input enables the first duplex S2S model without requiring speech pretrain. Separate architectures for agent and user modeling facilitate codec fine-tuning for better agent voices and halve the bitrate (0.6 kbps) compared to previous works. Experimental results show that the proposed model outperforms previous duplex models in reasoning, turn-taking, and barge-in abilities. The model requires significantly less speech data, as speech pretrain is skipped, which markedly simplifies the process of building a duplex S2S model from any LLMs. Finally, it is the first openly available duplex S2S model with training and inference code to foster reproducibility.

cs.CL↗

arXiv2025

High-Harmonic Generation and Optical Torque Interaction via Relativistic Diffraction of a Spatiotemporal Vortex Light

Diffraction of a relativistically-strong light can produce high-order harmonics via the relativistic oscillating window mechanism. In this process, the characteristics of the 2D electron dynamics at the diffraction screen can be imprinted to the generated harmonics, which provides abundant opportunities for manipulating light-matter interaction. In this work, we study single-slit diffraction of a high-intensity spatiotemporal optical vortex (STOV) - a beam carrying transverse orbital angular momentum (OAM). We demonstrate that due to the spatiotemporal structure of the driver, it induces differential electron oscillations on the screen, which conveys the transverse OAM to the high-order harmonic STOV beams. As a result, the topological charges of the harmonic beams are $l = nl_0$, where $l_0$ is the topological charge of the fundamental driving light, and $n$ is the harmonic order. In addition, we show that by controlling the slit angle with respect to the driver' s transverse OAM, the STOV light can exert a torque on the plasma, thus providing a way to manipulate the transverse OAM orientation of the generated harmonics.

physics.plasm-ph↗

arXiv2025

Training and Inference Efficiency of Encoder-Decoder Speech Models

Attention encoder-decoder model architecture is the backbone of several recent top performing foundation speech models: Whisper, Seamless, OWSM, and Canary-1B. However, the reported data and compute requirements for their training are prohibitive for many in the research community. In this work, we focus on the efficiency angle and ask the questions of whether we are training these speech models efficiently, and what can we do to improve? We argue that a major, if not the most severe, detrimental factor for training efficiency is related to the sampling strategy of sequential data. We show that negligence in mini-batch sampling leads to more than 50% computation being spent on padding. To that end, we study, profile, and optimize Canary-1B training to show gradual improvement in GPU utilization leading up to 5x increase in average batch sizes versus its original training settings. This in turn allows us to train an equivalent model using 4x less GPUs in the same wall time, or leverage the original resources and train it in 2x shorter wall time. Finally, we observe that the major inference bottleneck lies in the autoregressive decoder steps. We find that adjusting the model architecture to transfer model parameters from the decoder to the encoder results in a 3x inference speedup as measured by inverse real-time factor (RTFx) while preserving the accuracy and compute requirements for convergence. The training code and models will be available as open-source.

cs.CL↗