arXiv subjects

Yifan Zhou

Publications and source records attributed to Yifan Zhou.

At least 91 records · Page 5Linked to original sources

arXiv2025

CryptoMoE: Privacy-Preserving and Scalable Mixture of Experts Inference via Balanced Expert Routing

Private large language model (LLM) inference based on cryptographic primitives offers a promising path towards privacy-preserving deep learning. However, existing frameworks only support dense LLMs like LLaMA-1 and struggle to scale to mixture-of-experts (MoE) architectures. The key challenge comes from securely evaluating the dynamic routing mechanism in MoE layers, which may reveal sensitive input information if not fully protected. In this paper, we propose CryptoMoE, the first framework that enables private, efficient, and accurate inference for MoE-based models. CryptoMoE balances expert loads to protect expert routing information and proposes novel protocols for secure expert dispatch and combine. CryptoMoE also develops a confidence-aware token selection strategy and a batch matrix multiplication protocol to improve accuracy and efficiency further. Extensive experiments on DeepSeekMoE-16.4B, OLMoE-6.9B, and QWenMoE-14.3B show that CryptoMoE achieves $2.8\sim3.5\times$ end-to-end latency reduction and $2.9\sim4.3\times$ communication reduction over a dense baseline with minimum accuracy loss. We also adapt CipherPrune (ICLR'25) for MoE inference and demonstrate CryptoMoE can reduce the communication by up to $4.3 \times$. Code is available at: https://github.com/PKU-SEC-Lab/CryptoMoE.

cs.CR↗

arXiv2025

Sensitivity to Sub-Io-sized Exosatellite Transits in the MIRI LRS Lightcurve of the Nearest Substellar Worlds

JWST's unprecedented sensitivity enables precise spectrophotometric monitoring of substellar worlds, revealing atmospheric variability driven by mechanisms operating across different pressure levels. This same precision now permits exceptionally sensitive searches for transiting exosatellites, small terrestrial companions to these worlds. Using a novel simultaneous dual-band search method to address host variability, we present a search for transiting exosatellites in an 8-hour JWST/MIRI LRS lightcurve of the nearby ($2.0\,pc$) substellar binary WISE J1049-5319AB, composed of two $\sim30 M_{\rm Jup}$ brown dwarfs separated by $3.5\,au$ and viewed near edge-on. Although we detect no statistically significant transits, our injection-recovery tests demonstrate sensitivity to satellites as small as $0.275\,R_{\oplus}$ ($0.96\,R_{\rm Io}$ or $\sim$1 lunar radius), corresponding to 300ppm transit depths, and satellite-to-host mass ratios $>$$10^{-6}$. This approach paves the way for detecting Galilean-moon analogs around directly imaged brown dwarfs, free-floating planets, and wide-orbit exoplanets, dozens of which are already scheduled for JWST lightcurve monitoring. In our Solar System, each giant planet hosts on average 3.5 moons above this threshold, suggesting that JWST now probes a regime where such companions are expected to be abundant. The technique and sensitivities demonstrated here mark a critical step toward detecting exosatellites and ultimately enabling constraints on the occurrence rates of small terrestrial worlds orbiting $1\text{-}70$$M_{\rm Jup}$ hosts.

astro-ph.EP↗

arXiv2025

MM-HELIX: Boosting Multimodal Long-Chain Reflective Reasoning with Holistic Platform and Adaptive Hybrid Policy Optimization

While current Multimodal Large Language Models (MLLMs) have demonstrated proficiency in reasoning tasks such as mathematics and logic, their capacity for long-chain reflective reasoning, a prerequisite for solving complex real-world problems, remains largely underexplored. In this work, we first conduct an extensive empirical investigation to evaluate this capability. Leveraging a carefully designed data synthesis engine, we construct MM-HELIX, a multimodal benchmark consisting 1,260 samples of 42 challenging synthetic tasks that require iterative thinking and backtracking. Empirical results on this benchmark reveal that existing MLLMs exhibit significant performance deficits in long-chain reflective reasoning. To address this limitation, we generate post-training data and further explore learning paradigms for exploiting such data. We first develop the Step-Elicited Response Generation pipeline to create MM-HELIX-100K, a large-scale dataset of 100k high-quality, reflective reasoning traces for instruction-tuning stage. Given that standard Reinforcement Learning fails on complex tasks due to sparse reward signals and catastrophic forgetting after Supervised Fine-Tuning, we propose Adaptive Hybrid Policy Optimization (AHPO), a novel training strategy that dynamically unifies offline supervision and online optimization into a single stage. This strategy enables the model to learn from expert data when rewards are sparse and conduct independent exploration once proficient. When applied to the Qwen2.5-VL-7B baseline, our method achieves a +18.6\% accuracy improvement on MM-HELIX benchmark and demonstrates strong generalization with a +5.7\% average performance gain on general mathematic and logic tasks. Our work demonstrate that reflective reasoning in MLLMs can be effectively learned and generalized, paving the way for developing more capable MLLMs.

cs.CV↗

arXiv2025

CLAQS: Compact Learnable All-Quantum Token Mixer with Shared-ansatz for Text Classification

Quantum compute is scaling fast, from cloud QPUs to high throughput GPU simulators, making it timely to prototype quantum NLP beyond toy tasks. However, devices remain qubit limited and depth limited, training can be unstable, and classical attention is compute and memory heavy. This motivates compact, phase aware quantum token mixers that stabilize amplitudes and scale to long sequences. We present CLAQS, a compact, fully quantum token mixer for text classification that jointly learns complex-valued mixing and nonlinear transformations within a unified quantum circuit. To enable stable end-to-end optimization, we apply l1 normalization to regulate amplitude scaling and introduce a two-stage parameterized quantum architecture that decouples shared token embeddings from a window-level quantum feed-forward module. Operating under a sliding-window regime with document-level aggregation, CLAQS requires only eight data qubits and shallow circuits, yet achieves 91.64% accuracy on SST-2 and 87.08% on IMDB, outperforming both classical Transformer baselines and strong hybrid quantum-classical counterparts.

quant-ph↗

arXiv2025

BaTCAVe: Trustworthy Explanations for Robot Behaviors

Black box neural networks are an indispensable part of modern robots. Nevertheless, deploying such high-stakes systems in real-world scenarios poses significant challenges when the stakeholders, such as engineers and legislative bodies, lack insights into the neural networks' decision-making process. Presently, explainable AI is primarily tailored to natural language processing and computer vision, falling short in two critical aspects when applied in robots: grounding in decision-making tasks and the ability to assess trustworthiness of their explanations. In this paper, we introduce a trustworthy explainable robotics technique based on human-interpretable, high-level concepts that attribute to the decisions made by the neural network. Our proposed technique provides explanations with associated uncertainty scores for the explanation by matching neural network's activations with human-interpretable visualizations. To validate our approach, we conducted a series of experiments with various simulated and real-world robot decision-making models, demonstrating the effectiveness of the proposed approach as a post-hoc, human-friendly robot diagnostic tool.

cs.RO↗

arXiv2025

InfoMosaic-Bench: Evaluating Multi-Source Information Seeking in Tool-Augmented Agents

Information seeking is a fundamental requirement for humans. However, existing LLM agents rely heavily on open-web search, which exposes two fundamental weaknesses: online content is noisy and unreliable, and many real-world tasks require precise, domain-specific knowledge unavailable from the web. The emergence of the Model Context Protocol (MCP) now allows agents to interface with thousands of specialized tools, seemingly resolving this limitation. Yet it remains unclear whether agents can effectively leverage such tools -- and more importantly, whether they can integrate them with general-purpose search to solve complex tasks. Therefore, we introduce InfoMosaic-Bench, the first benchmark dedicated to multi-source information seeking in tool-augmented agents. Covering six representative domains (medicine, finance, maps, video, web, and multi-domain integration), InfoMosaic-Bench requires agents to combine general-purpose search with domain-specific tools. Tasks are synthesized with InfoMosaic-Flow, a scalable pipeline that grounds task conditions in verified tool outputs, enforces cross-source dependencies, and filters out shortcut cases solvable by trivial lookup. This design guarantees both reliability and non-triviality. Experiments with 14 state-of-the-art LLM agents reveal three findings: (i) web information alone is insufficient, with GPT-5 achieving only 38.2% accuracy and 67.5% pass rate; (ii) domain tools provide selective but inconsistent benefits, improving some domains while degrading others; and (iii) 22.4% of failures arise from incorrect tool usage or selection, highlighting that current LLMs still struggle with even basic tool handling.

cs.CL↗

arXiv2025

Fingerprinting Organic Molecules for the Inverse Design of Two-Dimensional Hybrid Perovskites with Target Energetics

Artificial intelligence (AI)-assisted workflows have transformed materials discovery, enabling rapid exploration of chemical spaces of functional materials. Endowed with extraordinary optoelectronic properties, two-dimensional (2D) hybrid perovskites represent an exciting frontier, but current efforts to design 2D perovskites rely heavily on trial-and-error and expert intuition approaches, leaving most of the chemical space unexplored and compromising the design of hybrid materials with desired properties. Here, we introduce an inverse design workflow for Dion-Jacobson perovskites that is built on an invertible fingerprint representation for millions of conjugated diammonium organic spacers. By incorporating high-throughput density functional theory (DFT) calculations, interpretable machine learning, and synthesis feasibility screening, we identified new organic spacer candidates with deterministic energy level alignment between the organic and the inorganic motifs in the 2D hybrid perovskites. These results highlight the power of integrating invertible, physically meaningful molecular representations into AI-assisted design, streamlining the property-targeted design of hybrid materials.

cond-mat.mtrl-sci↗

arXiv2025

The JWST Weather Report: retrieving temperature variations, auroral heating, and static cloud coverage on SIMP-0136

SIMP-0136 is a T2.5 brown dwarf whose young age ($200\pm50$~Myr) and low mass ($15\pm3$~M$_{\rm Jup}$) make it an ideal analogue for the directly imaged exoplanet population. With a 2.4 hour period, it is known to be variable in both the infrared and the radio, which has been attributed to changes in the cloud coverage and the presence of an aurora respectively. To quantify the changes in the atmospheric state that drive this variability, we obtained time-series spectra of SIMP-0136 covering one full rotation with both NIRSpec/PRISM and the MIRI/LRS on board JWST. We performed a series of time-resolved atmospheric retrievals using petitRADTRANS in order to measure changes in the temperature structure, chemistry, and cloudiness. We inferred the presence of a ~250 K thermal inversion above 10 mbar of SIMP-0136 at all phases, and propose that this inversion is due to the deposition of energy into the upper atmosphere by an aurora. Statistical tests were performed in order to determine which parameters drive the observed spectroscopic variability. The primary contribution was due to changes in the temperature profile at pressures deeper than 10 mbar, which resulted in variation of the effective temperature from 1243 K to 1248 K. This changing effective temperature was also correlated to observed changes in the abundances of CO2 and H2S, while all other chemical species were consistent with being homogeneous throughout the atmosphere. Patchy silicate clouds were required to fit the observed spectra, but the cloud properties were not found to systematically vary with longitude. This work paints a portrait of an L/T transition object where the primary variability mechanisms are magnetic and thermodynamic in nature, rather than due to inhomogeneous cloud coverage.

astro-ph.EP↗

arXiv2025

Alias-Free Latent Diffusion Models: Improving Fractional Shift Equivariance of Diffusion Latent Space

Latent Diffusion Models (LDMs) are known to have an unstable generation process, where even small perturbations or shifts in the input noise can lead to significantly different outputs. This hinders their applicability in applications requiring consistent results. In this work, we redesign LDMs to enhance consistency by making them shift-equivariant. While introducing anti-aliasing operations can partially improve shift-equivariance, significant aliasing and inconsistency persist due to the unique challenges in LDMs, including 1) aliasing amplification during VAE training and multiple U-Net inferences, and 2) self-attention modules that inherently lack shift-equivariance. To address these issues, we redesign the attention modules to be shift-equivariant and propose an equivariance loss that effectively suppresses the frequency bandwidth of the features in the continuous domain. The resulting alias-free LDM (AF-LDM) achieves strong shift-equivariance and is also robust to irregular warping. Extensive experiments demonstrate that AF-LDM produces significantly more consistent results than vanilla LDM across various applications, including video editing and image-to-image translation.

cs.CV↗

arXiv2025

Characterizing the Time Variability of 2M1207 A+b with JWST NIRSpec/PRISM

We present JWST NIRSpec/PRISM IFU time-resolved observations of 2M1207 A and b (TWA 27), a $\sim 10$ Myr binary system consisting of a $\sim 2500$ K sub-stellar primary hosting a $\sim 1300$ K companion. Our data provide 20 time-resolved spectra over an observation spanning 12.56 hours. We provide an empirical characterization for the spectra of both objects across time. For 2M1207 A, non-linear trend models are statistically favored within the ranges 0.6-2.3 $μ$m and 3.8-5.3 $μ$m. However, most of the periods constrained from sinusoidal models exceed the observing window, setting a lower limit of 12.56 hours. We find the data at H$α$ and beyond 4.35 $μ$m show a moderate time correlation, as well as a pair of light curves at 0.73-0.80 $μ$m and 3.36-3.38 $μ$m. For 2M1207 b, light curves integrated across 0.86-1.77 $μ$m and 3.29-4.34 $μ$m support linear trend models. Following the interpretation of Zhang et. al. (2025), we model the 2M1207 b data with two 1D atmospheric components, both with silicate and iron condensates. The model of time variability as changes to the cloud filling factor shows broad consistency with the variability amplitudes derived from our data. Our amplitudes, however, disagree with the models at $\approx$0.86-1 $μ$m. While an additional model component such as rainout chemistry may be considered here, our analysis is limited by a low signal-to-noise ratio. Our results demonstrate the capability of JWST to simultaneously monitor the spectral variability of a planetary-mass companion and host at low contrast.

astro-ph.EP↗

arXiv2025

Seismic Oscillations Excited by Giant Impacts in Directly-Imaged Giant Planets

The radii and masses of many giant exoplanets imply their interiors each contain more than $\sim$100 $M_\oplus$ of solids. A large metal content may arise when a giant planet grows by colliding and merging with multiple $\sim$10 $M_\oplus$ solid cores. Here we show that a giant impact with a young gas giant excites long-lived seismic oscillations that can be detected photometrically. Mode lifetimes are close to the planet's Kelvin-Helmholtz time, a significant fraction of a young planet's age. Oscillation periods lie between tens of minutes to an hour, and variability amplitudes can exceed a percent for several million years. Beta Pictoris b is a young super-Jupiter known to be highly metal-enriched. If a Neptune-mass (17 $M_\oplus$) body impacted $β$ Pictoris b in the past $\sim$9--18 Myr, the planet could still be ringing with a percent-level photometric variability measurable with JWST.

astro-ph.EP↗

arXiv2025

Laplacian Analysis Meets Dynamics Modelling: Gaussian Splatting for 4D Reconstruction

While 3D Gaussian Splatting (3DGS) excels in static scene modeling, its extension to dynamic scenes introduces significant challenges. Existing dynamic 3DGS methods suffer from either over-smoothing due to low-rank decomposition or feature collision from high-dimensional grid sampling. This is because of the inherent spectral conflicts between preserving motion details and maintaining deformation consistency at different frequency. To address these challenges, we propose a novel dynamic 3DGS framework with hybrid explicit-implicit functions. Our approach contains three key innovations: a spectral-aware Laplacian encoding architecture which merges Hash encoding and Laplacian-based module for flexible frequency motion control, an enhanced Gaussian dynamics attribute that compensates for photometric distortions caused by geometric deformation, and an adaptive Gaussian split strategy guided by KDTree-based primitive control to efficiently query and optimize dynamic areas. Through extensive experiments, our method demonstrates state-of-the-art performance in reconstructing complex dynamic scenes, achieving better reconstruction fidelity.

cs.GR↗

arXiv2025

Alignment and Safety in Large Language Models: Safety Mechanisms, Training Paradigms, and Emerging Challenges

Due to the remarkable capabilities and growing impact of large language models (LLMs), they have been deeply integrated into many aspects of society. Thus, ensuring their alignment with human values and intentions has emerged as a critical challenge. This survey provides a comprehensive overview of practical alignment techniques, training protocols, and empirical findings in LLM alignment. We analyze the development of alignment methods across diverse paradigms, characterizing the fundamental trade-offs between core alignment objectives. Our analysis shows that while supervised fine-tuning enables basic instruction-following, preference-based methods offer more flexibility for aligning with nuanced human intent. We discuss state-of-the-art techniques, including Direct Preference Optimization (DPO), Constitutional AI, brain-inspired methods, and alignment uncertainty quantification (AUQ), highlighting their approaches to balancing quality and efficiency. We review existing evaluation frameworks and benchmarking datasets, emphasizing limitations such as reward misspecification, distributional robustness, and scalable oversight. We summarize strategies adopted by leading AI labs to illustrate the current state of practice. We conclude by outlining open problems in oversight, value pluralism, robustness, and continuous alignment. This survey aims to inform both researchers and practitioners navigating the evolving landscape of LLM alignment.

cs.AI↗

arXiv2025

Outdoor Monocular SLAM with Global Scale-Consistent 3D Gaussian Pointmaps

3D Gaussian Splatting (3DGS) has become a popular solution in SLAM due to its high-fidelity and real-time novel view synthesis performance. However, some previous 3DGS SLAM methods employ a differentiable rendering pipeline for tracking, lack geometric priors in outdoor scenes. Other approaches introduce separate tracking modules, but they accumulate errors with significant camera movement, leading to scale drift. To address these challenges, we propose a robust RGB-only outdoor 3DGS SLAM method: S3PO-GS. Technically, we establish a self-consistent tracking module anchored in the 3DGS pointmap, which avoids cumulative scale drift and achieves more precise and robust tracking with fewer iterations. Additionally, we design a patch-based pointmap dynamic mapping module, which introduces geometric priors while avoiding scale ambiguity. This significantly enhances tracking accuracy and the quality of scene reconstruction, making it particularly suitable for complex outdoor environments. Our experiments on the Waymo, KITTI, and DL3DV datasets demonstrate that S3PO-GS achieves state-of-the-art results in novel view synthesis and outperforms other 3DGS SLAM methods in tracking accuracy. Project page: https://3dagentworld.github.io/S3PO-GS/.

cs.CV↗

arXiv2025

X-Intelligence 3.0: Training and Evaluating Reasoning LLM for Semiconductor Display

Large language models (LLMs) have recently achieved significant advances in reasoning and demonstrated their advantages in solving challenging problems. Yet, their effectiveness in the semiconductor display industry remains limited due to a lack of domain-specific training and expertise. To bridge this gap, we present X-Intelligence 3.0, the first high-performance reasoning model specifically developed for the semiconductor display industry. This model is designed to deliver expert-level understanding and reasoning for the industry's complex challenges. Leveraging a carefully curated industry knowledge base, the model undergoes supervised fine-tuning and reinforcement learning to enhance its reasoning and comprehension capabilities. To further accelerate development, we implemented an automated evaluation framework that simulates expert-level assessments. We also integrated a domain-specific retrieval-augmented generation (RAG) mechanism, resulting in notable performance gains on benchmark datasets. Despite its relatively compact size of 32 billion parameters, X-Intelligence 3.0 outperforms SOTA DeepSeek-R1-671B across multiple evaluations. This demonstrates its exceptional efficiency and establishes it as a powerful solution to the longstanding reasoning challenges faced by the semiconductor display industry.

cs.CL↗

arXiv2025

Wavelet-GS: 3D Gaussian Splatting with Wavelet Decomposition

3D Gaussian Splatting (3DGS) has revolutionized 3D scene reconstruction, which effectively balances rendering quality, efficiency, and speed. However, existing 3DGS approaches usually generate plausible outputs and face significant challenges in complex scene reconstruction, manifesting as incomplete holistic structural outlines and unclear local lighting effects. To address these issues simultaneously, we propose a novel decoupled optimization framework, which integrates wavelet decomposition into 3D Gaussian Splatting and 2D sampling. Technically, through 3D wavelet decomposition, our approach divides point clouds into high-frequency and low-frequency components, enabling targeted optimization for each. The low-frequency component captures global structural outlines and manages the distribution of Gaussians through voxelization. In contrast, the high-frequency component restores intricate geometric and textural details while incorporating a relight module to mitigate lighting artifacts and enhance photorealistic rendering. Additionally, a 2D wavelet decomposition is applied to the training images, simulating radiance variations. This provides critical guidance for high-frequency detail reconstruction, ensuring seamless integration of details with the global structure. Extensive experiments on challenging datasets demonstrate our method achieves state-of-the-art performance across various metrics, surpassing existing approaches and advancing the field of 3D scene reconstruction.

cs.GR↗

arXiv2025

TokensGen: Harnessing Condensed Tokens for Long Video Generation

Generating consistent long videos is a complex challenge: while diffusion-based generative models generate visually impressive short clips, extending them to longer durations often leads to memory bottlenecks and long-term inconsistency. In this paper, we propose TokensGen, a novel two-stage framework that leverages condensed tokens to address these issues. Our method decomposes long video generation into three core tasks: (1) inner-clip semantic control, (2) long-term consistency control, and (3) inter-clip smooth transition. First, we train To2V (Token-to-Video), a short video diffusion model guided by text and video tokens, with a Video Tokenizer that condenses short clips into semantically rich tokens. Second, we introduce T2To (Text-to-Token), a video token diffusion transformer that generates all tokens at once, ensuring global consistency across clips. Finally, during inference, an adaptive FIFO-Diffusion strategy seamlessly connects adjacent clips, reducing boundary artifacts and enhancing smooth transitions. Experimental results demonstrate that our approach significantly enhances long-term temporal and content coherence without incurring prohibitive computational overhead. By leveraging condensed tokens and pre-trained short video models, our method provides a scalable, modular solution for long video generation, opening new possibilities for storytelling, cinematic production, and immersive simulations. Please see our project page at https://vicky0522.github.io/tokensgen-webpage/ .

cs.CV↗

arXiv2025

Towards Efficient Privacy-Preserving Machine Learning: A Systematic Review from Protocol, Model, and System Perspectives

Privacy-preserving machine learning (PPML) based on cryptographic protocols has emerged as a promising paradigm to protect user data privacy in cloud-based machine learning services. While it achieves formal privacy protection, PPML often incurs significant efficiency and scalability costs due to orders of magnitude overhead compared to the plaintext counterpart. Therefore, there has been a considerable focus on mitigating the efficiency gap for PPML. In this survey, we provide a comprehensive and systematic review of recent PPML studies with a focus on cross-level optimizations. Specifically, we categorize existing papers into protocol level, model level, and system level, and review progress at each level. We also provide qualitative and quantitative comparisons of existing works with technical insights, based on which we discuss future research directions and highlight the necessity of integrating optimizations across protocol, model, and system levels. We hope this survey can provide an overarching understanding of existing approaches and potentially inspire future breakthroughs in the PPML field. As the field is evolving fast, we also provide a public GitHub repository to continuously track the developments, which is available at https://github.com/PKU-SEC-Lab/Awesome-PPML-Papers.

cs.CR↗