Searcharxiv⌕ Search

arXiv subjects

Xiaodong Shi

Publications and source records attributed to Xiaodong Shi.

At least 19 recordsLinked to original sources
arXiv2026

CHIME: Credit-Aware Hierarchical Memory Evolution for Long-Horizon Agentic Planning

Planning is a central capability that enables agents to decompose complex long-horizon tasks into manageable steps. Test-time search and training-based methods improve planning but incur high inference costs or require expensive training data. Self-evolving memory instead accumulates reusable experience from agent interaction outcomes into an external memory bank, so planning capability keeps improving at inference time without parameter updates. However, existing self-evolving memory methods share an inherent credit assignment problem: they rely on final task outcomes as feedback, but such outcomes conflate plan quality with execution errors and environmental factors, so the accumulated planning experience is often biased and noisy. To address this problem, we propose Credit-Aware Hierarchical Memory Evolution (CHIME), a self-evolving memory framework that maintains a separate planning bank and execution bank and follows an attribute-before-memorize principle: CHIME first attributes each task outcome to the plan, the execution, both, or neither, and then updates only the corresponding memory bank. Extensive experiments on four long-horizon agent benchmarks show that CHIME consistently outperforms state-of-the-art training-based and self-evolving memory baselines. Further analyses reveal several interesting findings. For example, CHIME accumulates effective memory with far fewer items. In addition, the learned memory values faithfully reflect downstream utility: high-quality planning memories are more valuable than execution memories. Finally, the accumulated memory effectively transfers across backbone models. Code will be released at https://github.com/ATH-MaaS/Marco-DeepResearch.

cs.AI↗
arXiv2026

MAG: MAnifold Guided Semi-Supervised Multi-modal In-Context Learning

Few-shot in-context learning (ICL) with multi-modal large language models (MLLMs) enables task adaptation without parameter updates, but its performance is highly sensitive to the quality and coverage of the selected demonstrations. While unlabeled multi-modal data is abundant, it remains elusive how to exploit them for ICL. We propose MAG (MAnifold-Guided semi-supervised in-context demonstra- tion selection), an efficient framework that leverages unlabeled data to improve multi-modal ICL. MAG formulates demonstration selection as a semi-supervised propagation problem on a multi-modal graph and adopts a two-stage strategy: (i) relevance score propagation identifies a compact set of high-impact unlabeled samples for pseudo-labeling, reducing MLLM inference cost; (ii) multi-modal relevance is used to select the final demonstrations. We show that textual represen- tations are more effective for relevance propagation, while both visual and textual modalities are crucial for high-quality demonstration selection. Experiments on eight multi-modal benchmarks demonstrate that MAG consistently outperforms strong baselines in label-scarce regimes, achieving significant gains with a limited pseudo-labeling budget.

cs.LG↗
arXiv2026

TARL: Transaction-Aware Reliable Ledgers for Executable Memory Management in Long-Term Agents

Persistent memory helps long-term agents retain knowledge, yet a single update error can repeatedly distort future retrieval and reasoning. Most existing systems reduce memory updating to a binary Write/Hold decision, which cannot distinguish whether new information should be added, ignored, used to revise an outdated belief, rejected as unreliable, or deferred for verification. These choices may share the same binary label while producing fundamentally different memory states. We introduce TARL, a memory state update framework that maps each statement to one of five executable actions. TARL identifies the affected memory, resolves its temporal scope, compares source reliability, and updates accepted, pending, and rejected ledgers. It is further trained by comparing the memory states produced by alternative update operations, encouraging the model to select the operation that leads to the correct result. We also introduce TARL-Mem, a benchmark with fine-grained action labels and next-state targets. Across in-domain, cross-source, temporal, counterfactual, and sequential evaluations, TARL improves action prediction and state recovery, reduces memory pollution, preserves conflicting evidence, and limits cumulative corruption.

cs.AI↗
arXiv2026

Integrated polarization-entangled photon source for wavelength-multiplexed quantum networks

Entangled photons are fundamental resources for quantum communication, computing, and networking. Among them, polarization-entangled photon pairs play an important role due to their straightforward state manipulation and direct use in quantum key distribution, teleportation, and network protocols. However, realizing compact, efficient, and scalable polarization-entangled sources that meet the requirements of practical deployment remains a major challenge. Here, we present a simple yet high-performance on-chip polarization-entangled photon-pair source on thin-film lithium niobate (TFLN). Our device employs dual quasi-phase matching (D-QPM) that sequentially supports type-0 and type-I spontaneous parametric down-conversion in a single nanophotonic waveguide, eliminating the need for interferometers, polarization rotators, or other complex circuits. The source directly produces high-fidelity Bell states with broad bandwidth, high brightness, and low noise. Using this integrated platform, we realize wavelength-multiplexed entanglement distribution in a four-user quantum network deployed over metropolitan fiber links up to 50 km. These results establish a robust and scalable pathway toward practical quantum communication systems and multi-user quantum mesh networks based on integrated photonics.

physics.optics↗
arXiv2026

PAMT: Process-Aligned Reinforcement Learning for Multi-Domain Machine Translation

Multi-domain machine translation (MDMT) requires more than fluent generation: it demands domain-sensitive translation decisions such as domain disambiguation, terminology control, and stylistic adaptation. Large reasoning models (LRMs) make such decisions explicit through intermediate translation steps, but our analysis across 15 domains and four translation directions shows that this explicit reasoning is double-edged: it improves long-form and high-difficulty translation, yet often drifts in terminology-intensive and stylistically constrained settings. We trace this failure to a credit-assignment bottleneck: existing methods optimize final outputs or coarse trajectories, but cannot identify which translation steps actually help the final translation. To address this, we propose PAMT, a process-aligned training framework that combines cold-start domain-aware Long-CoT supervision with reinforcement learning. PAMT uses sequence-level format and outcome rewards for the final translation, together with a step-level process reward that measures how much each explicit translation step increases the likelihood of the reference translation. Across two backbones, PAMT improves over base models, outperforms MT-specialized baselines on average, and remains competitive with strong LLMs/LRMs across in-domain, OOD, and multilingual settings.

cs.CL↗
arXiv2026

Don't Peek at the Answer: Outcome-Masked Group Relative Policy Optimization for Label-Free RLVR

Reinforcement Learning with Verifiable Rewards (RLVR) improves LLM reasoning but typically relies on ground-truth (GT) answers, limiting scalability. Voting-based label-free RLVR replace gold supervision with answer-level consensus from model samples. However, collapse arises when the same answer-level signal is used both to estimate rewards and to drive token-level policy optimization, encouraging the model to directly reinforce answer tokens rather than improve reasoning. We propose OM-GRPO, a label-free RLVR framework that decouples reward estimation from policy optimization. OM-GRPO masks gradients on the answer span while retaining answer-level rewards through a soft consensus signal, shifting optimization pressure away from answer tokens. We further introduce Contrast-Augmented Reward, which refines reward estimation via low-cost pairwise comparisons over existing trajectories without additional rollouts. Across diverse reasoning benchmarks and three LLM backbones, OM-GRPO consistently outperforms existing label-free RLVR methods and matches supervised GT-reward training with stable optimization. This stability is particularly beneficial in the Test-Time Training setting, where OM-GRPO surpasses majority voting by 4.24 points.

cs.AI↗
arXiv2026

Translation with Thought: Difficulty-Adaptive Reasoning via Reinforcement Learning for Multi-Domain Machine Translation

Multi-domain machine translation (MDMT) poses a unique challenge due to varying levels of linguistic complexity across domains. Inspired by human translators' ability to adapt reasoning effort based on difficulty, we propose TwT (Translation with Thought), a resource-rational framework that learns to modulate inference between intuitive and deliberate reasoning. TwT is trained in two stages: (1) supervised fine-tuning on difficulty-aware long chain-of-thought traces distilled from DeepSeek-R1 and rewritten by GPT-4o to reflect human-like reasoning economy, and (2) reinforcement learning with a hybrid reward to optimize translation quality and reasoning efficiency. Evaluated on 15 benchmarks spanning in-domain and out-of-domain settings, as well as 3 seen and 59 unseen languages, with ablations across three backbone models, TwT-7B and TwT-14B outperform much larger SOTA reasoning models in translation quality, while reducing token usage by 32--60\%. These results confirm that aligning translation behavior with cognitive principles enables robust generalization, high translation quality, and efficient reasoning in MDMT.

cs.CL↗
arXiv2026

Segment-chirped periodically poled lithium niobate waveguides for broadband supercontinuum generation

Supercontinuum generation is a key technology in nonlinear optics, supporting a wide range of applications in frequency metrology and spectroscopy. Integrated photonics offers a promising route toward compact and efficient supercontinuum sources, yet extending the bandwidth while maintaining high spectral flatness remains a central challenge. Here we demonstrate an integrated broadband supercontinuum source based on segment-chirped periodically poled lithium niobate (SC-PPLN) nanophotonic waveguides. By discretizing the chirped poling profile into independently optimized segments, this approach enables high-fidelity ferroelectric domain inversion with near-ideal duty cycles and establishes broadband quasi-phase matching, overcoming the domain inhomogeneity and efficiency limitations commonly encountered in conventional chirped poling. The engineered phase-matching landscape supports efficient wavelength conversion and simultaneous activation of multiple second- and third-order nonlinear processes. Experimentally, we achieve a spectrally flat supercontinuum spanning three optical octaves, from 320 nm in the ultraviolet to 2600 nm in the mid-infrared. These results establish segment-chirped poling as a practical strategy for broadband wavelength conversion and supercontinuum generation in integrated photonics.

physics.optics↗
arXiv2026

Integrated photon-pair sources on periodically poled thin-film lithium tantalate

Chip-integrated photon-pair sources based on spontaneous parametric down-conversion (SPDC) have emerged as a promising solution for scalable quantum light generation. Thin-film lithium tantalate (TFLT) is a compelling $χ^{(2)}$ platform, combining strong nonlinearity with a high optical-damage threshold, weak photorefractive response, and ferroelectricity that enables quasi-phase matching. However, SPDC-based photon-pair generation on TFLT has not yet been demonstrated. Here, we combine high-quality periodic poling with low-loss nanophotonic waveguides to realize photon-pair sources on TFLT in both traveling-wave and resonant configurations. In periodically poled straight waveguides, we achieve broadband photon-pair generation with high efficiency ($2.1~\mathrm{GHz}~\mathrm{mW}^{-1}$) and coincidence-to-accidental ratio (up to $3.8\times10^{5}$). We further confirm high-purity single-photon operation via heralded second-order correlation ($g^{(2)}_\mathrm{H}(0) = 0.0018 \pm 0.0002$) and high-fidelity time-energy entanglement through Franson interference (visibility of $98.9 \pm 0.5\%$). In periodically poled racetrack resonators, we map out a broad quantum frequency comb spanning the telecom C- and L-bands. By isolating individual frequency-correlated pairs, we measure a high spectral brightness of $11~\mathrm{GHz}~\mathrm{mW}^{-1}~\mathrm{GHz}^{-1}$. These results are competitive with the state of the art across $χ^{(2)}$ integrated platforms, positioning TFLT as a strong contender for integrated quantum light sources, with applications in wavelength-multiplexed quantum communications and photonic quantum information processing.

physics.optics↗
arXiv2026

Passive All-Optical Nonlinear Neuron Activation via PPLN Nanophotonic Waveguides

Artificial intelligence (AI) is transforming modern life, yet the growing scale of AI applications places mounting demands on computational resources, raising sustainability concerns. Photonic integrated circuits (PICs) offer a promising alternative, enabling massive parallelism, low latency, and reduced electrical overhead, particularly excelling in high-throughput linear operations. However, passive and fully optical nonlinear activation functions with equally superb performance remain rare, posing a critical bottleneck in realizing all-optical neural networks in PICs. Here, we demonstrate a compact and integrated all-optical nonlinear activation method, experimentally realized through strong second-order optical nonlinearities in periodically poled lithium niobate (PPLN) nanophotonic waveguides, achieving 80% absolute conversion efficiency. This activation exhibits a sigmoid-like, wavelength-selective response with femtosecond-scale dynamics and light-speed processing, operating passively without external control and auxiliary signals. We validate its feasibility for neural inference by cascading the PPLN-driven activations with a linear silicon PIC, demonstrating all-optical nonlinear neuron expressivity. Moreover, combining the measured nonlinearity with linear operations calculated by the PIC, we show that PPLN-activated multi-layer optical neural networks can achieve performance on par with digital implementations in real-world tasks, including airfoil regression and medical image classification. These results pave the way toward scalable, high-speed, and fully integrated all-optical neural networks for next-generation photonic AI hardware.

physics.optics↗
arXiv2026

UMEM: Unified Memory Extraction and Management Framework for Generalizable Memory

Self-evolving memory serves as the trainable parameters for Large Language Models (LLMs)-based agents, where extraction (distilling insights from experience) and management (updating the memory bank) must be tightly coordinated. Existing methods predominately optimize memory management while treating memory extraction as a static process, resulting in poor generalization, where agents accumulate instance-specific noise rather than robust memories. To address this, we propose Unified Memory Extraction and Management (UMEM), a self-evolving agent framework that jointly optimizes a Large Language Model to simultaneous extract and manage memories. To mitigate overfitting to specific instances, we introduce Semantic Neighborhood Modeling and optimize the model with a neighborhood-level marginal utility reward via GRPO. This approach ensures memory generalizability by evaluating memory utility across clusters of semantically related queries. Extensive experiments across five benchmarks demonstrate that UMEM significantly outperforms highly competitive baselines, achieving up to a 10.67% improvement in multi-turn interactive tasks. Futhermore, UMEM maintains a monotonic growth curve during continuous evolution. Codes and models will be publicly released.

cs.CL↗
arXiv2026

DARL: Encouraging Diverse Answers for General Reasoning without Verifiers

Reinforcement Learning with Verifiable Rewards (RLVR) has demonstrated promising gains in enhancing the reasoning capabilities of large language models. However, its dependence on domain-specific verifiers significantly restricts its applicability to open and general domains. Recent efforts such as RLPR have extended RLVR to general domains, enabling training on broader datasets and achieving improvements over RLVR. However, a notable limitation of these methods is their tendency to overfit to reference answers, which constrains the model's ability to generate diverse outputs. This limitation is particularly pronounced in open-ended tasks such as writing, where multiple plausible answers exist. To address this, we propose DARL, a simple yet effective reinforcement learning framework that encourages the generation of diverse answers within a controlled deviation range from the reference while preserving alignment with it. Our framework is fully compatible with existing general reinforcement learning methods and can be seamlessly integrated without additional verifiers. Extensive experiments on thirteen benchmarks demonstrate consistent improvements in reasoning performance. Notably, DARL surpasses RLPR, achieving average gains of 1.3 points on six reasoning benchmarks and 9.5 points on seven general benchmarks, highlighting its effectiveness in improving both reasoning accuracy and output diversity.

cs.CL↗
arXiv2026

On-Chip Generation of Co-Polarized and Spectrally Separable Photon Pairs

On-chip generation of high-purity single photons is essential for scalable photonic quantum technologies. Spontaneous parametric down-conversion (SPDC) is widely used to generate photon pairs for heralded single-photon sources, but intrinsic spectral correlations of the pairs often limit the purity and interference visibility of the heralded photons. Existing approaches to suppress these correlations rely on narrowband spectral filtering, which introduces loss, or exploiting different polarizations, which complicates on-chip integration. Here, we demonstrate a new strategy for generating spectrally separable photon pairs in thin-film lithium niobate nanophotonic circuits by harnessing higher-order spatial modes, with all interacting fields residing in the same polarization. Spectral separability is achieved by engineering group-velocity matching using higher-order transverse-electric modes, combined with a Gaussian-apodized poling profile to further suppress residual correlations inherent to standard periodic poling. Subsequent on-chip mode conversion with efficiency exceeding 95\% maps the higher-order mode to the fundamental mode and routes the photons into distinct output channels. The resulting heralded photons exhibit spectral purities exceeding 94\% inferred from joint-spectral intensity and 89\% from unheralded $g^{(2)}$ measurement. This approach enables flexible spectral and temporal engineering of on-chip quantum light sources for quantum computing and quantum networking.

quant-ph↗
arXiv2025

Squeezed Light Generation in Periodically Poled Thin-Film Lithium Niobate Waveguides

Squeezed states of light play a key role in quantum-enhanced sensing and continuous-variable quantum information processing. Realizing integrated squeezed light sources is crucial for developing compact and scalable photonic quantum systems. In this work, we demonstrate on-chip broadband vacuum squeezing at telecommunication wavelengths on the thin-film lithium niobate (TFLN) platform. Our device integrates periodically poled lithium niobate (PPLN) nanophotonic waveguides with low-loss edge couplers, comprising bilayer inverse tapers and an SU-8 polymer waveguide. This configuration achieves a fiber-to-chip coupling loss of 1.4 dB and a total homodyne detection loss of 4 dB, enabling a measured squeezing level of 1.4 dB. Additional measurements in a more efficient PPLN waveguide (without low-loss couplers) infer an on-chip squeezing level of over 10 dB at a pump power of 62 mW. These results underscore the potential of TFLN platform for efficient and scalable squeezed light generation.

physics.optics↗
arXiv2025

PerformSinger: Multimodal Singing Voice Synthesis Leveraging Synchronized Lip Cues from Singing Performance Videos

Existing singing voice synthesis (SVS) models largely rely on fine-grained, phoneme-level durations, which limits their practical application. These methods overlook the complementary role of visual information in duration prediction.To address these issues, we propose PerformSinger, a pioneering multimodal SVS framework, which incorporates lip cues from video as a visual modality, enabling high-quality "duration-free" singing voice synthesis. PerformSinger comprises parallel multi-branch multimodal encoders, a feature fusion module, a duration and variational prediction network, a mel-spectrogram decoder and a vocoder. The fusion module, composed of adapter and fusion blocks, employs a progressive fusion strategy within an aligned semantic space to produce high-quality multimodal feature representations, thereby enabling accurate duration prediction and high-fidelity audio synthesis. To facilitate the research, we design, collect and annotate a novel SVS dataset involving synchronized video streams and precise phoneme-level manual annotations. Extensive experiments demonstrate the state-of-the-art performance of our proposal in both subjective and objective evaluations. The code and dataset will be publicly available.

eess.AS↗
arXiv2025

From Neurons to Semantics: Evaluating Cross-Linguistic Alignment Capabilities of Large Language Models via Neurons Alignment

Large language models (LLMs) have demonstrated remarkable multilingual capabilities, however, how to evaluate cross-lingual alignment remains underexplored. Existing alignment benchmarks primarily focus on sentence embeddings, but prior research has shown that neural models tend to induce a non-smooth representation space, which impact of semantic alignment evaluation on low-resource languages. Inspired by neuroscientific findings that similar information activates overlapping neuronal regions, we propose a novel Neuron State-Based Cross-Lingual Alignment (NeuronXA) to assess the cross-lingual a lignment capabilities of LLMs, which offers a more semantically grounded approach to assess cross-lingual alignment. We evaluate NeuronXA on several prominent multilingual LLMs (LLaMA, Qwen, Mistral, GLM, and OLMo) across two transfer tasks and three multilingual benchmarks. The results demonstrate that with only 100 parallel sentence pairs, NeuronXA achieves a Pearson correlation of 0.9556 with downstream tasks performance and 0.8514 with transferability. These findings demonstrate NeuronXA's effectiveness in assessing both cross-lingual alignment and transferability, even with a small dataset. This highlights its potential to advance cross-lingual alignment research and to improve the semantic understanding of multilingual LLMs.

cs.CL↗
arXiv2025

LLMs Can Achieve High-quality Simultaneous Machine Translation as Efficiently as Offline

When the complete source sentence is provided, Large Language Models (LLMs) perform excellently in offline machine translation even with a simple prompt "Translate the following sentence from [src lang] into [tgt lang]:". However, in many real scenarios, the source tokens arrive in a streaming manner and simultaneous machine translation (SiMT) is required, then the efficiency and performance of decoder-only LLMs are significantly limited by their auto-regressive nature. To enable LLMs to achieve high-quality SiMT as efficiently as offline translation, we propose a novel paradigm that includes constructing supervised fine-tuning (SFT) data for SiMT, along with new training and inference strategies. To replicate the token input/output stream in SiMT, the source and target tokens are rearranged into an interleaved sequence, separated by special tokens according to varying latency requirements. This enables powerful LLMs to learn read and write operations adaptively, based on varying latency prompts, while still maintaining efficient auto-regressive decoding. Experimental results show that, even with limited SFT data, our approach achieves state-of-the-art performance across various SiMT benchmarks, and preserves the original abilities of offline translation. Moreover, our approach generalizes well to document-level SiMT setting without requiring specific fine-tuning, even beyond the offline translation model.

cs.CL↗
arXiv2025

How Well Do Large Reasoning Models Translate? A Comprehensive Evaluation for Multi-Domain Machine Translation

Large language models (LLMs) have demonstrated strong performance in general-purpose machine translation, but their effectiveness in complex, domain-sensitive translation tasks remains underexplored. Recent advancements in Large Reasoning Models (LRMs), raise the question of whether structured reasoning can enhance translation quality across diverse domains. In this work, we compare the performance of LRMs with traditional LLMs across 15 representative domains and four translation directions. Our evaluation considers various factors, including task difficulty, input length, and terminology density. We use a combination of automatic metrics and an enhanced MQM-based evaluation hierarchy to assess translation quality. Our findings show that LRMs consistently outperform traditional LLMs in semantically complex domains, especially in long-text and high-difficulty translation scenarios. Moreover, domain-adaptive prompting strategies further improve performance by better leveraging the reasoning capabilities of LRMs. These results highlight the potential of structured reasoning in MDMT tasks and provide valuable insights for optimizing translation systems in domain-sensitive contexts.

cs.CL↗