Source author record

Yuchen Zhang

Yuchen Zhang appears in the imported research catalog. Authorship, coauthor and topic links are available while profile ownership is still unclaimed.

ResearcherUnclaimed source record

Catalog footprint

What is connected

41works

24topics

4close collaborators

Actions

Connect this record

Open graph Browse works

Inspect adjacent papers, topics, institutions and collaborators without losing the researcher page.

Building this map preview

BZPEER is loading the nearby papers, people, topics and institutions for this page.

preprint2026arXiv

6D Movable Antenna Enhanced Cell-free MIMO: Two-timescale Decentralized Beamforming and Antenna Movement Optimization

This paper investigates a six-dimensional movable antenna (6DMA)-aided cell-free multi-user multiple-input multiple-output (MIMO) communication system. In this system, each distributed access point (AP) can flexibly adjust its array orientation and antenna positions to adapt to spatial channel variations and enhance communication performance. However, frequent antenna movements and centralized beamforming based on global instantaneous channel state information (CSI) sharing among APs entail extremely high signal processing delay and system overhead, which is difficult to be practically implemented in high-mobility scenarios with short channel coherence time. To address these practical implementation challenges and improve scalability, a two-timescale decentralized optimization framework is proposed in this paper to jointly design the beamformer, antenna positions, and array orientations. In the short timescale, each AP updates its receive beamformer based on local instantaneous CSI and global statistical CSI. In the long timescale, the central processing unit optimizes the antenna positions and array orientations at all APs based on global statistical CSI to maximize the ergodic sum rate of all users. The resulting optimization problem is non-convex and involves highly coupled variables, thus posing significant challenges for obtaining efficient solutions. To address this problem, a constrained stochastic successive convex approximation algorithm is developed. Numerical results demonstrate that the proposed 6DMA-aided cell-free system with decentralized beamforming significantly outperforms other antenna movement schemes with less flexibility and even achieves a performance comparable to that of the centralized beamforming benchmark.

preprint2026arXiv

Beyond Known Objects: A Novel Framework for Open-Set Object Detection using Negative-Aware Norm

Open-Set Object Detection (OSOD) is crucial for autonomous driving, where perception systems must recognize and localize both known and previously unseen objects in complex, dynamic environments. While recent approaches deliver promising results, they often require retraining the detector extensively to learn objectness, which describes the likelihood that a bounding box tightly encloses a valid object, regardless of whether its category was learned during training. Deviating from existing work, we hypothesize that standard off-the-shelf detectors may already contain helpful cues for objectness, owing to their training on numerous and diverse known categories. Building on this idea, we propose NAN-SPOT, a training-light framework that does not require to retrain the base object detector and estimates objectness by leveraging a hidden layer metric called Negative-Aware Norm (NAN), requiring only minutes of training on just hundreds of images. To support comprehensive evaluation, we introduce COCO-Open, an expanded version of the existing COCO-Mixed dataset, increasing unknown object annotations from 433 to 1853, making it the most exhaustively labeled dataset for OSOD to the best of our knowledge. Experimental results demonstrate that NAN-SPOT achieves even better performance on unknown object detection than methods requiring heavy training, without compromising performance on known objects. This efficiency and robustness make NAN-SPOT a promising step towards open-world perception in autonomous driving.

preprint2026arXiv

Characterizing the Consistency of the Emergent Misalignment Persona

Fine-tuning large language models (LLMs) on narrowly misaligned data generalizes to broadly misaligned behavior, a phenomenon termed emergent misalignment (EM). While prior work has found a correlation between harmful behavior and self-assessment in emergently misaligned models, it remains unclear how consistent this correspondence is across tasks and whether it varies across fine-tuning domains. We characterize the consistency of the EM persona by fine-tuning Qwen 2.5 32B Instruct on six narrowly misaligned domains (e.g., insecure code, risky financial advice, bad medical advice) and administering experiments including harmfulness evaluation, self-assessment, choosing between two descriptions of AI systems, output recognition, and score prediction. Our results reveal two distinct patterns: coherent-persona models, in which harmful behavior and self-reported misalignment are coupled, and inverted-persona models, which produce harmful outputs while identifying as aligned AI systems. These findings reveal a more fine-grained picture of the effects of emergent misalignment, calling into question the consistency of the EM persona.

preprint2026arXiv

Continuous Discovery of Vulnerabilities in LLM Serving Systems with Fuzzing

LLM inference and serving systems have become security-critical infrastructure; however, many of their most concerning failures arise from the serving layer rather than from model behavior alone. Modern inference engines combine KV cache, batching, prefix sharing, speculative decoding, adapters, and multi-tenant scheduling, creating shared-state behavior that only emerges under realistic concurrent workloads and is missed by standard model, safety, and API tests. We present GRIEF, a greybox fuzzer for LLM inference engines that treats timed multi-request traces as first-class inputs, uses lightweight oracles to detect crashes, hangs, performance pathologies, and silent output corruption, and applies controlled replay with log-probability checks to confirm reproducible serving-layer failures. Across early campaigns on vLLM and SGLang, GRIEF discovers 15 vulnerabilities, 10 confirmed by engine developers, including 2 CVEs, spanning KV-cache isolation failures, cross-request performance interference, and crash or liveness bugs. These results show that concurrency, caching, and state reuse can induce silent cross-request contamination, noisy-neighbor denial of service, and delayed crashes without malformed inputs or explicit server errors, making concurrent serving behavior a first-class security and reliability boundary for LLM infrastructure.

preprint2026arXiv

Distributed Integrated Sensing, Localization, and Communications over LEO Satellite Constellations

Low Earth orbit (LEO) satellite constellations are rapidly becoming essential enablers of next-generation wireless systems, offering global broadband access, high-precision localization, and reliable sensing beyond terrestrial coverage. However, the inherent limitations of individual LEO satellites, including restricted power, limited antenna aperture, and constrained onboard processing, hinder their ability to meet the growing demands of 6G applications. To address these challenges, this article introduces the concept of distributed integrated sensing, localization, and communication (DISLAC) over LEO constellations, inspired by distributed multiple input multiple output architectures. By enabling inter-satellite cooperation through inter-satellite links, DISLAC jointly exploits communication, localization, and sensing functionalities, achieving synergistic gains in throughput, positioning accuracy, and sensing robustness through shared resources and cooperative design. We present illustrative case studies that quantify these benefits and analyze key system-level considerations, including synchronization, antenna reconfigurability, and inter-satellite link design. The article concludes by outlining open research directions to advance the practical deployment of DISLAC in future non-terrestrial networks.

preprint2026arXiv

DynaTrain: Fast Online Parallelism Switching for Elastic LLM Training

Modern large language model (LLM) training is inherently dynamic: resource fluctuations, RLHF phase shifts, and cluster elasticity continually reshape the optimal parallelism layout, posing a significant challenge to existing training frameworks built around a static execution model. We present DynaTrain, a distributed training system for sub-second, online reconfiguration across arbitrary multi-dimensional parallelism. At its core, we propose a Virtual Parameter Space (VPS) abstraction that unifies all distributed training states under one logical coordinate space, turning any parallelism configuration into a deterministic mapping and collapsing complex transition into manageable geometric intersections. On top of VPS, a state routing-and-transition layer executes rank-local transfers under a memory-aware, deadlock-free schedule, and an Elastic Device Manager overlaps new-world construction with ongoing training to mask topology-change cost. On dense and MoE models up to 235B parameters, DynaTrain reconfigures a 70B dense model in under 2s and a 235B MoE model in 4.36s, outperforming state-of-the-art checkpoint-based and elastic systems by up to three orders of magnitude while preserving correctness.

preprint2026arXiv

GESR: Graph-Based Edge Semantic Reconstruction for Stealthy Communication Detection with Benign-Only Training

Detecting stealthy malicious communications from flow logs under benign-only training remains a critical challenge in network security. Malicious communications often camouflage as normal traffic like standard HTTPS flows. Conventional intrusion detectors rely strictly on known labeled attacks. Alternatively, they score flows completely independently. These approaches fail against sparse and context-dependent suspicious activity. To capture this essential context, graph anomaly detectors have been introduced to add valuable relational information to the analysis. However, existing methods fail to test the structural consistency of specific communication edges. To overcome these fundamental limitations, we present GESR, a novel graph-based framework for detecting suspicious communications and anomalous hosts under a benign-only training setting. GESR models complex network activity as attributed communication graphs. It cleverly reconstructs edge semantics entirely from local structural context rather than isolated features. This non-intuitive design forces the framework to predict expected communication patterns from neighborhood topologies. Attackers cannot easily manipulate this deep structural dependency. The model then converts the resulting structural inconsistencies into host-level anomaly scores. It utilizes robust Median Absolute Deviation (MAD) calibration for this final step. We evaluate GESR extensively on CTU-13 and CICIDS2017 datasets. These evaluations strictly impose tight false-positive operating constraints. On CICIDS2017, GESR achieves an outstanding ROC-AUC of 0.9753. It also yields a high TPR of 0.8569 at a strict 5% FPR threshold. GESR consistently outperforms existing methods across both evaluated benchmarks. The results prove that structure-conditioned edge reconstruction is a credible direction for practical intrusion detection.

preprint2026arXiv

Implicit Compression Regularization: Concise Reasoning via Internal Shorter Distributions in RL Post-Training

Reinforcement learning with verifiable rewards improves LLM reasoning but often induces overthinking, where models generate unnecessarily long reasoning traces. Existing methods mainly rely on length penalties or early-exit strategies; however, the former may degrade accuracy and induce underthinking, whereas the latter assumes that substantial portions of reasoning traces can be safely truncated. To obtain a compression signal without these limitations, we revisit the training dynamics of existing compression methods. We observe that the length--accuracy correlation is initially negative but continually increases during compression, indicating that shorter responses are initially more likely to be correct but gradually lose this property as the policy moves toward underthinking. Based on this observation, we formalize overthinking: a negative correlation indicates an overthinking regime, while a positive one indicates underthinking. When overthinking, the shortest correct responses are shorter than the group-average response length in expectation, making them natural compression targets already present in on-policy rollouts. We therefore propose \emph{Implicit Compression Regularization} (ICR), an on-policy regularization method whose compression signal comes from a virtual shorter distribution induced by the shortest correct responses in rollout groups, guiding the policy toward concise yet correct trajectories. Training dynamics show that ICR maintains a better length--accuracy correlation during compression, indicating that short responses remain better aligned with correctness instead of drifting toward underthinking. Experiments on three reasoning backbones and multiple mathematical and knowledge-intensive benchmarks show that ICR consistently shortens responses while preserving or improving accuracy, achieving a stronger accuracy--length Pareto frontier.

preprint2026arXiv

MT-JailBench: A Modular Benchmark for Understanding Multi-Turn Jailbreak Attacks

Multi-turn jailbreaks exploit the ability of large language models to accumulate and act on conversational context. Instead of stating a harmful request directly, an attacker can gradually steer the conversation toward an unsafe answer. Recent methods demonstrate this risk, but they are usually evaluated as black-box pipelines with different budgets, judges, retry rules, and strategy generation procedures. As a result, it is often unclear whether reported gains reflect stronger attack mechanisms or different experimental conditions. We introduce MT-JailBench, a modular evaluation framework for benchmarking multi-turn jailbreaks under fixed conditions. MT-JailBench implements each attack as five interacting modules: evaluation function, attack strategy, prompt generation, prompt refinement, and flow control. This design enables fair comparison across attack methods and component-wise analysis of what drives attack success. Using MT-JailBench, we find that resource budgets and evaluation functions are major confounders: controlling turns, retries, interactions, sampled strategies, and judges substantially change the ranking of attacks. At the component level, prompt generation accounts for most performance variation, while refinement and flow control provide moderate gains. We also find that explicit dynamic strategy generation is not always necessary; stochastic sampling from a fixed strategy can rival more elaborate diversification mechanisms. Finally, recomposing the best components yields a strong attack configuration that outperforms its source attacks and generalizes across diverse target LLMs. MT-JailBench therefore provides a modular framework for comparing multi-turn jailbreaks, understanding the impact of components, and guiding stronger red-teaming evaluations.

preprint2026arXiv

Nonlinear virtual lens for programmable and multispectral infrared upconversion imaging

Conventional infrared (IR) imaging techniques depend on IR cameras based on narrow-bandgap semiconductors, which offer limited spectral bandwidth, coupled with a separate lens. Recently, advances in nonlinear flat optics have opened a novel pathway for converting IR signals into the visible through nonlinear generations, enabling the direct visualisation of IR images using standard visible cameras. However, the narrow spectral bandwidth and the requirement for an additional lens remain the key challenges. Here, we address both issues via a novel adaptive and multifunctional IR-to-visible imaging platform offering tunable bandwidth and focusing simultaneously. We utilise sum-frequency generation (SFG) to convert IR light into the visible, by introducing a pump beam modulated by a spatial light modulator (SLM) to construct a virtual metalens enabling precisely controlled focusing of the generated nonlinear optical field. As a result, we demonstrate both theoretically and experimentally an optical focusing mechanism with a tunable focal length, achieved by varying the pump and signal wavelengths and modulating the phase distribution. Furthermore, since the focal length depends on the input signal wavelength, the imaging plane position varies accordingly, indicating a promising potential for the multispectral IR imaging applications. Our upconversion platform delivers SLM-controlled, programmable multispectral focusing for next-generation IR imaging, opening new avenues in the fields of computational and multispectral imaging techniques.

preprint2026arXiv

OmniVL-Guard Pro: A Tool-Augmented Agent for Omnibus Vision-Language Forensics

Existing vision-language forgery detection and grounding methods operate under a closed-world paradigm, assuming verification can be completed by the model alone. However, self-contained MLLMs are constrained by finite parametric knowledge, static training corpora, and limited perceptual resolution, creating a practical ceiling in dynamic open-world forensics -- particularly for real-time event verification requiring external clues and forgery segmentation demanding fine-grained scrutiny of local manipulations. To address these limitations, we shift from scaling up the self-contained model toward reaching beyond it. We propose \textbf{OmniVL-Guard Pro}, a tool-augmented agent that extends unified forensics from closed-world prediction to open-world clues-driven reasoning. OmniVL-Guard Pro integrates a tool environment spanning real-time event search, local cropping and zooming, edge-anomaly screening, face detection, video frame extraction, and SAM3-based segmentation. To generate high-quality tool-reasoning trajectories, we introduce \textbf{Tree-Structured Self-Evolving Tool Trajectory Generation}, which produces diverse trajectories through seed guidance, guider-free self-evolution, and weakly-hinted hard sample synthesis, yielding the Full-Spectrum Tool Reasoning (FSTR) dataset for training. We further propose \textbf{Checker-Guided Agentic Reinforcement Learning} (CGARL), which provides process-level supervision to penalize cases where the answer is correct but the reasoning is distorted. Extensive experiments demonstrate that OmniVL-Guard Pro achieves state-of-the-art performance across various tasks, and exhibits strong zero-shot generalization. The FSTR dataset and code for OmniVL-Guard Pro will be publicly released at \url{https://github.com/shen8424/OmniVL-Guard-Pro}.

preprint2026arXiv

Positioning-Aided Channel Estimation for Multi-LEO Satellite Cooperative Beamforming

We investigate a multi-low Earth orbit (LEO) satellite system that simultaneously provides positioning and communication services to terrestrial user terminals. To address the challenges of accurately acquiring channel state information in LEO satellite systems, we propose a novel two-timescale positioning-aided channel estimation framework, exploiting the distinct variation rates of position-related parameters and channel gains inherent in LEO satellite channels. Using the misspecified Cramér-Rao bound (MCRB) theory, we systematically analyze positioning performance under practical imperfections, such as inter-satellite clock bias and carrier frequency offset. Furthermore, we theoretically demonstrate how position information derived from downlink positioning can enhance uplink channel estimation accuracy, even in the presence of positioning errors, through an MCRB-based analysis. To address the limited link budgets and communication rates of single-satellite communication, we develop a multi-LEO cooperative beamforming strategy for downlink transmission that leverages cluster-wise satellite cooperation while maintaining reduced complexity. Theoretical analyses and numerical results confirm the effectiveness of the proposed framework in facilitating high-precision downlink positioning under practical imperfections, facilitating uplink channel estimation, and enabling efficient downlink communication.

preprint2026arXiv

Precoding Matrix Indicator in the 5G NR Protocol: A Tutorial on 3GPP Beamforming Codebooks

This paper bridges this critical gap by providing a systematic examination of the beamforming codebook technology, i.e., precoding matrix indicator (PMI), in the 5G NR from theoretical, standardization, and implementation perspectives. We begin by introducing the background of beamforming in multiple-input multiple-output (MIMO) systems and the signaling procedures for codebook-based beamforming in practical 5G systems. Then, we establish the fundamentals of regular codebooks and port-selection codebooks in 3GPP standards. Next, we provide rigorous technical analysis of 3GPP codebook evolution spanning Releases 15-18, with particular focus on: 1) We elucidate the core principles underlying codebook design, 2) provide clear physical interpretations for each symbolic variable in the codebook formulas, summarized in tabular form, and 3) offer intuitive visual illustrations to explain how codebook parameters convey information. These essential pedagogical elements are almost entirely absent in the often-obscure standardization documents. Through mathematical modeling, performance benchmarking, feedback comparisons, and scenario-dependent applicability analysis, we provide researchers and engineers with a unified understanding of beamforming codebooks in real-world systems. Furthermore, we identify future directions and other beamforming scenarios for ongoing research and development efforts. This work serves as both an informative tutorial and a guidance for future research, facilitating more effective collaboration between academia and industry in advancing wireless communication technologies.

preprint2026arXiv

Teaching Thinking Models to Reason with Tools: A Full-Pipeline Recipe for Tool-Integrated Reasoning

Tool-integrated reasoning (TIR) offers a direct way to extend thinking models beyond the limits of text-only reasoning. Paradoxically, we observe that tool-enabled evaluation can degrade reasoning performance even when the strong thinking models make almost no actual tool calls. In this paper, we investigate how to inject natural tool-use behavior into a strong thinking model without sacrificing its no-tool reasoning ability, and present a comprehensive TIR recipe. We highlight that (i) the effectiveness of TIR supervised fine-tuning (SFT) hinges on the learnability of teacher trajectories, which should prioritize problems inherently suited for tool-augmented solutions; (ii) controlling the proportion of tool-use trajectories could mitigate the catastrophic forgetting of text-only reasoning capacity; (iii) optimizing for pass@k and response length instead of training loss could maximize TIR SFT gains while preserving headroom for reinforcement learning (RL) exploration; (iv) a stable RL with verifiable rewards (RLVR) stage, built upon suitable SFT initialization and explicit safeguards against mode collapse, provides a simple yet remarkably effective solution. When applied to Qwen3 thinking models at 4B and 30B scales, our recipe yields models that achieve state-of-the-art performance in a wide range of benchmarks among open-source models, such as 96.7% and 99.2% on AIME 2025 for 4B and 30B, respectively.

preprint2026arXiv

The RoboSense Challenge: Sense Anything, Navigate Anywhere, Adapt Across Platforms

Autonomous systems are increasingly deployed in open and dynamic environments -- from city streets to aerial and indoor spaces -- where perception models must remain reliable under sensor noise, environmental variation, and platform shifts. However, even state-of-the-art methods often degrade under unseen conditions, highlighting the need for robust and generalizable robot sensing. The RoboSense 2025 Challenge is designed to advance robustness and adaptability in robot perception across diverse sensing scenarios. It unifies five complementary research tracks spanning language-grounded decision making, socially compliant navigation, sensor configuration generalization, cross-view and cross-modal correspondence, and cross-platform 3D perception. Together, these tasks form a comprehensive benchmark for evaluating real-world sensing reliability under domain shifts, sensor failures, and platform discrepancies. RoboSense 2025 provides standardized datasets, baseline models, and unified evaluation protocols, enabling large-scale and reproducible comparison of robust perception methods. The challenge attracted 143 teams from 85 institutions across 16 countries, reflecting broad community engagement. By consolidating insights from 23 winning solutions, this report highlights emerging methodological trends, shared design principles, and open challenges across all tracks, marking a step toward building robots that can sense reliably, act robustly, and adapt across platforms in real-world environments.

preprint2023arXiv

Mining User-aware Multi-relations for Fake News Detection in Large Scale Online Social Networks

Users' involvement in creating and propagating news is a vital aspect of fake news detection in online social networks. Intuitively, credible users are more likely to share trustworthy news, while untrusted users have a higher probability of spreading untrustworthy news. In this paper, we construct a dual-layer graph (i.e., the news layer and the user layer) to extract multiple relations of news and users in social networks to derive rich information for detecting fake news. Based on the dual-layer graph, we propose a fake news detection model named Us-DeFake. It learns the propagation features of news in the news layer and the interaction features of users in the user layer. Through the inter-layer in the graph, Us-DeFake fuses the user signals that contain credibility information into the news features, to provide distinctive user-aware embeddings of news for fake news detection. The training process conducts on multiple dual-layer subgraphs obtained by a graph sampler to scale Us-DeFake in large scale social networks. Extensive experiments on real-world datasets illustrate the superiority of Us-DeFake which outperforms all baselines, and the users' credibility signals learned by interaction relation can notably improve the performance of our model.

preprint2022arXiv

Building Embedded Systems Like It's 1996

Embedded devices are ubiquitous. However, preliminary evidence shows that attack mitigations protecting our desktops/servers/phones are missing in embedded devices, posing a significant threat to embedded security. To this end, this paper presents an in-depth study on the adoption of common attack mitigations on embedded devices. Precisely, it measures the presence of standard mitigations against memory corruptions in over 10k Linux-based firmware of deployed embedded devices. The study reveals that embedded devices largely omit both user-space and kernel-level attack mitigations. The adoption rates on embedded devices are multiple times lower than their desktop counterparts. An equally important observation is that the situation is not improving over time. Without changing the current practices, the attack mitigations will remain missing, which may become a bigger threat in the upcoming IoT era. Throughout follow-up analyses, we further inferred a set of factors possibly contributing to the absence of attack mitigations. The exemplary ones include massive reuse of non-protected software, lateness in upgrading outdated kernels, and restrictions imposed by automated building tools. We envision these will turn into insights towards improving the adoption of attack mitigations on embedded devices in the future.

preprint2022arXiv

Distributed Estimation for Interconnected Systems with Arbitrary Coupling Structures

This paper is concerned with the problem of distributed estimation for time-varying interconnected dynamic systems with arbitrary coupling structures. To guarantee the robustness of the designed estimators, novel distributed stability conditions are proposed with only local information and the information from neighbors. Then, simplified stability conditions which do not require timely exchange of neighbors' estimator gain information is further developed for systems with delayed communication. By merging these subsystem-level stability conditions and the optimization-based estimator gain design, the distributed, stable and optimal estimators are proposed. Quite notably, these optimization solutions can be easily obtained by standard software packages, and it is also shown that the designed estimators are scalable in the sense of adding or subtracting subsystems. Finally, an illustrative example is employed to show the effectiveness of the proposed methods.

preprint2022arXiv

Performance Analysis for Covert Communications Under Faster-than-Nyquist Signaling

In this letter, we analyze the performance of covert communications under faster-than-Nyquist (FTN) signaling in the Rayleigh block fading channel. Both Bayesian criterion- and Kullback-Leibler (KL) divergence-based covertness constraints are considered. Especially, for KL divergence-based one, we prove that both the maximum transmit power and covert rate under FTN signaling are higher than those under Nyquist signaling. Numerical results coincide with our analysis and validate the advantages of FTN signaling to realize covert data transmission.

preprint2021arXiv

Adversarial example generation with AdaBelief Optimizer and Crop Invariance

Deep neural networks are vulnerable to adversarial examples, which are crafted by applying small, human-imperceptible perturbations on the original images, so as to mislead deep neural networks to output inaccurate predictions. Adversarial attacks can thus be an important method to evaluate and select robust models in safety-critical applications. However, under the challenging black-box setting, most existing adversarial attacks often achieve relatively low success rates on adversarially trained networks and advanced defense models. In this paper, we propose AdaBelief Iterative Fast Gradient Method (ABI-FGM) and Crop-Invariant attack Method (CIM) to improves the transferability of adversarial examples. ABI-FGM and CIM can be readily integrated to build a strong gradient-based attack to further boost the success rates of adversarial examples for black-box attacks. Moreover, our method can also be naturally combined with other gradient-based attack methods to build a more robust attack to generate more transferable adversarial examples against the defense models. Extensive experiments on the ImageNet dataset demonstrate the method's effectiveness. Whether on adversarially trained networks or advanced defense models, our method has higher success rates than state-of-the-art gradient-based attack methods.

preprint2021arXiv

Task-Oriented Dialogue as Dataflow Synthesis

We describe an approach to task-oriented dialogue in which dialogue state is represented as a dataflow graph. A dialogue agent maps each user utterance to a program that extends this graph. Programs include metacomputation operators for reference and revision that reuse dataflow fragments from previous turns. Our graph-based state enables the expression and manipulation of complex user intents, and explicit metacomputation makes these intents easier for learned models to predict. We introduce a new dataset, SMCalFlow, featuring complex dialogues about events, weather, places, and people. Experiments show that dataflow graphs and metacomputation substantially improve representability and predictability in these natural dialogues. Additional experiments on the MultiWOZ dataset show that our dataflow representation enables an otherwise off-the-shelf sequence-to-sequence model to match the best existing task-specific state tracking model. The SMCalFlow dataset and code for replicating experiments are available at https://www.microsoft.com/en-us/research/project/dataflow-based-dialogue-semantic-machines.

preprint2020arXiv

Chandra Detection of Three X-ray Bright Quasars at z>5

We report Chandra detection of three UV bright radio quiet quasars at $z\gtrsim5$. We have collected a sufficient number of photons to extract an X-ray spectrum of each quasar to measure their basic X-ray properties, such as the X-ray flux, power law photon index ($Γ$), and optical-to-X-ray spectral slope ($α_{\rm OX}$). J074749+115352 at $z=5.26$ is the X-ray brightest radio-quiet quasar at $z>5$. It may have a short timescale variation (on a timescale of $\sim3800\rm~s$ in the observer's frame, or $\sim600\rm~s$ in the rest frame) which is however largely embedded in the statistical noise. We extract phase folded spectra of this quasar. There are two distinguishable states: a "high soft" state with an average X-ray flux $\sim2.7$ times of the "low hard" state, and a significantly steeper X-ray spectral slope ($Γ=2.40_{-0.32}^{+0.33}$ vs $1.78_{-0.24}^{+0.25}$). We also compare the three quasars detected in this paper to other quasar samples. We find that J074749+115352, with a SMBH mass of $M_{\rm SMBH}\approx1.8\times10^9\rm~M_\odot$ and an Eddington ratio of $λ_{\rm Edd}\approx2.3$, is extraordinarily X-ray bright. It has an average $α_{\rm OX}=-1.46\pm0.02$ and a 2-10 keV bolometric correction factor of $L_{\rm bol}/L_{\rm2-10keV}=42.4\pm5.8$, both significantly depart from some well defined scaling relations. We compare $Γ$ of the three quasars to other samples at different redshifts, and do not find any significant redshift evolution based on the limited sample of $z>5$ quasars with reliable measurements of the X-ray spectral properties.

preprint2020arXiv

On Localized Discrepancy for Domain Adaptation

We propose the discrepancy-based generalization theories for unsupervised domain adaptation. Previous theories introduced distribution discrepancies defined as the supremum over complete hypothesis space. The hypothesis space may contain hypotheses that lead to unnecessary overestimation of the risk bound. This paper studies the localized discrepancies defined on the hypothesis space after localization. First, we show that these discrepancies have desirable properties. They could be significantly smaller than the pervious discrepancies. Their values will be different if we exchange the two domains, thus can reveal asymmetric transfer difficulties. Next, we derive improved generalization bounds with these discrepancies. We show that the discrepancies could influence the rate of the sample complexity. Finally, we further extend the localized discrepancies for achieving super transfer and derive generalization bounds that could be even more sample-efficient on source domain.

preprint2017arXiv

On isolated singularities with noninvertible finite endomorphism

We prove that if $ϕ:(X,0)\to (X,0)$ is a finite endomorphism of an isolated singularity such that $\operatorname{deg}(ϕ)\geq 2$ and $ϕ$ is étale in codimension 1, then $X$ is $\mathbb{Q}$-Gorenstein and log canonical.

preprint2016arXiv

Convexified Convolutional Neural Networks

We describe the class of convexified convolutional neural networks (CCNNs), which capture the parameter sharing of convolutional neural networks in a convex manner. By representing the nonlinear convolutional filters as vectors in a reproducing kernel Hilbert space, the CNN parameters can be represented as a low-rank matrix, which can be relaxed to obtain a convex optimization problem. For learning two-layer convolutional neural networks, we prove that the generalization error obtained by a convexified CNN converges to that of the best possible CNN. For learning deeper networks, we train CCNNs in a layer-wise manner. Empirically, CCNNs achieve performance competitive with CNNs trained by backpropagation, SVMs, fully-connected neural networks, stacked denoising auto-encoders, and other baseline methods.

preprint2016arXiv

Local Maxima in the Likelihood of Gaussian Mixture Models: Structural Results and Algorithmic Consequences

We provide two fundamental results on the population (infinite-sample) likelihood function of Gaussian mixture models with $M \geq 3$ components. Our first main result shows that the population likelihood function has bad local maxima even in the special case of equally-weighted mixtures of well-separated and spherical Gaussians. We prove that the log-likelihood value of these bad local maxima can be arbitrarily worse than that of any global optimum, thereby resolving an open question of Srebro (2007). Our second main result shows that the EM algorithm (or a first-order variant of it) with random initialization will converge to bad critical points with probability at least $1-e^{-Ω(M)}$. We further establish that a first-order variant of EM will not converge to strict saddle points almost surely, indicating that the poor performance of the first-order method can be attributed to the existence of bad local maxima rather than bad saddle points. Overall, our results highlight the necessity of careful initialization when using the EM algorithm in practice, even when applied in highly favorable settings.

preprint2016arXiv

On Bayes Risk Lower Bounds

This paper provides a general technique for lower bounding the Bayes risk of statistical estimation, applicable to arbitrary loss functions and arbitrary prior distributions. A lower bound on the Bayes risk not only serves as a lower bound on the minimax risk, but also characterizes the fundamental limit of any estimator given the prior knowledge. Our bounds are based on the notion of $f$-informativity, which is a function of the underlying class of probability measures and the prior. Application of our bounds requires upper bounds on the $f$-informativity, thus we derive new upper bounds on $f$-informativity which often lead to tight Bayes risk lower bounds. Our technique leads to generalizations of a variety of classical minimax bounds (e.g., generalized Fano's inequality). Our Bayes risk lower bounds can be directly applied to several concrete estimation problems, including Gaussian location models, generalized linear models, and principal component analysis for spiked covariance models. To further demonstrate the applications of our Bayes risk lower bounds to machine learning problems, we present two new theoretical results: (1) a precise characterization of the minimax risk of learning spherical Gaussian mixture models under the smoothed analysis framework, and (2) lower bounds for the Bayes risk under a natural prior for both the prediction and estimation errors for high-dimensional sparse linear regression under an improper learning setting.

preprint2015arXiv

$\ell_1$-regularized Neural Networks are Improperly Learnable in Polynomial Time

We study the improper learning of multi-layer neural networks. Suppose that the neural network to be learned has $k$ hidden layers and that the $\ell_1$-norm of the incoming weights of any neuron is bounded by $L$. We present a kernel-based method, such that with probability at least $1 - δ$, it learns a predictor whose generalization error is at most $ε$ worse than that of the neural network. The sample complexity and the time complexity of the presented method are polynomial in the input dimension and in $(1/ε,\log(1/δ),F(k,L))$, where $F(k,L)$ is a function depending on $(k,L)$ and on the activation function, independent of the number of neurons. The algorithm applies to both sigmoid-like activation functions and ReLU-like activation functions. It implies that any sufficiently sparse neural network is learnable in polynomial time.

preprint2015arXiv

Communication-Efficient Distributed Optimization of Self-Concordant Empirical Loss

We consider distributed convex optimization problems originated from sample average approximation of stochastic optimization, or empirical risk minimization in machine learning. We assume that each machine in the distributed computing system has access to a local empirical loss function, constructed with i.i.d. data sampled from a common distribution. We propose a communication-efficient distributed algorithm to minimize the overall empirical loss, which is the average of the local empirical losses. The algorithm is based on an inexact damped Newton method, where the inexact Newton steps are computed by a distributed preconditioned conjugate gradient method. We analyze its iteration complexity and communication efficiency for minimizing self-concordant empirical loss functions, and discuss the results for distributed ridge regression, logistic regression and binary classification with a smoothed hinge loss. In a standard setting for supervised learning, the required number of communication rounds of the algorithm does not increase with the sample size, and only grows slowly with the number of machines.

preprint2015arXiv

Distributed Estimation of Generalized Matrix Rank: Efficient Algorithms and Lower Bounds

We study the following generalized matrix rank estimation problem: given an $n \times n$ matrix and a constant $c \geq 0$, estimate the number of eigenvalues that are greater than $c$. In the distributed setting, the matrix of interest is the sum of $m$ matrices held by separate machines. We show that any deterministic algorithm solving this problem must communicate $Ω(n^2)$ bits, which is order-equivalent to transmitting the whole matrix. In contrast, we propose a randomized algorithm that communicates only $\widetilde O(n)$ bits. The upper bound is matched by an $Ω(n)$ lower bound on the randomized communication complexity. We demonstrate the practical effectiveness of the proposed algorithm with some numerical experiments.

preprint2015arXiv

Learning Halfspaces and Neural Networks with Random Initialization

We study non-convex empirical risk minimization for learning halfspaces and neural networks. For loss functions that are $L$-Lipschitz continuous, we present algorithms to learn halfspaces and multi-layer neural networks that achieve arbitrarily small excess risk $ε>0$. The time complexity is polynomial in the input dimension $d$ and the sample size $n$, but exponential in the quantity $(L/ε^2)\log(L/ε)$. These algorithms run multiple rounds of random initialization followed by arbitrary optimization steps. We further show that if the data is separable by some neural network with constant margin $γ>0$, then there is a polynomial-time algorithm for learning a neural network that separates the training data with margin $Ω(γ)$. As a consequence, the algorithm achieves arbitrary generalization error $ε>0$ with ${\rm poly}(d,1/ε)$ sample and time complexity. We establish the same learnability result when the labels are randomly flipped with probability $η<1/2$.

preprint2015arXiv

Optimal prediction for sparse linear models? Lower bounds for coordinate-separable M-estimators

For the problem of high-dimensional sparse linear regression, it is known that an $\ell_0$-based estimator can achieve a $1/n$ "fast" rate on the prediction error without any conditions on the design matrix, whereas in absence of restrictive conditions on the design matrix, popular polynomial-time methods only guarantee the $1/\sqrt{n}$ "slow" rate. In this paper, we show that the slow rate is intrinsic to a broad class of M-estimators. In particular, for estimators based on minimizing a least-squares cost function together with a (possibly non-convex) coordinate-wise separable regularizer, there is always a "bad" local optimum such that the associated prediction error is lower bounded by a constant multiple of $1/\sqrt{n}$. For convex regularizers, this lower bound applies to all global optima. The theory is applicable to many popular estimators, including convex $\ell_1$-based methods as well as M-estimators based on nonconvex regularizers, including the SCAD penalty or the MCP regularizer. In addition, for a broad class of nonconvex regularizers, we show that the bad local optima are very common, in that a broad class of local minimization algorithms with random initialization will typically converge to a bad solution.

preprint2015arXiv

Splash: User-friendly Programming Interface for Parallelizing Stochastic Algorithms

Stochastic algorithms are efficient approaches to solving machine learning and optimization problems. In this paper, we propose a general framework called Splash for parallelizing stochastic algorithms on multi-node distributed systems. Splash consists of a programming interface and an execution engine. Using the programming interface, the user develops sequential stochastic algorithms without concerning any detail about distributed computing. The algorithm is then automatically parallelized by a communication-efficient execution engine. We provide theoretical justifications on the optimal rate of convergence for parallelizing stochastic gradient descent. Splash is built on top of Apache Spark. The real-data experiments on logistic regression, collaborative filtering and topic modeling verify that Splash yields order-of-magnitude speedup over single-thread stochastic algorithms and over state-of-the-art implementations on Spark.

preprint2015arXiv

Stochastic Primal-Dual Coordinate Method for Regularized Empirical Risk Minimization

We consider a generic convex optimization problem associated with regularized empirical risk minimization of linear predictors. The problem structure allows us to reformulate it as a convex-concave saddle point problem. We propose a stochastic primal-dual coordinate (SPDC) method, which alternates between maximizing over a randomly chosen dual variable and minimizing over the primal variable. An extrapolation step on the primal variable is performed to obtain accelerated convergence rate. We also develop a mini-batch version of the SPDC method which facilitates parallel computing, and an extension with weighted sampling probabilities on the dual variables, which has a better complexity than uniform sampling on unnormalized data. Both theoretically and empirically, we show that the SPDC method has comparable or better performance than several state-of-the-art optimization methods.

preprint2014arXiv

Divide and Conquer Kernel Ridge Regression: A Distributed Algorithm with Minimax Optimal Rates

We establish optimal convergence rates for a decomposition-based scalable approach to kernel ridge regression. The method is simple to describe: it randomly partitions a dataset of size N into m subsets of equal size, computes an independent kernel ridge regression estimator for each subset, then averages the local solutions into a global predictor. This partitioning leads to a substantial reduction in computation time versus the standard approach of performing kernel ridge regression on all N samples. Our two main theorems establish that despite the computational speed-up, statistical optimality is retained: as long as m is not too large, the partition-based estimator achieves the statistical minimax rate over all estimators using the set of N samples. As concrete examples, our theory guarantees that the number of processors m may grow nearly linearly for finite-rank kernels and Gaussian kernels and polynomially in N for Sobolev spaces, which in turn allows for substantial reductions in computational cost. We conclude with experiments on both simulated data and a music-prediction task that complement our theoretical results, exhibiting the computational and statistical benefits of our approach.

preprint2014arXiv

Lower bounds on the performance of polynomial-time algorithms for sparse linear regression

Under a standard assumption in complexity theory (NP not in P/poly), we demonstrate a gap between the minimax prediction risk for sparse linear regression that can be achieved by polynomial-time algorithms, and that achieved by optimal algorithms. In particular, when the design matrix is ill-conditioned, the minimax prediction loss achievable by polynomial-time algorithms can be substantially greater than that of an optimal algorithm. This result is the first known gap between polynomial and optimal algorithms for sparse linear regression, and does not depend on conjectures in average-case complexity.

preprint2014arXiv

On the generic vanishing theorem of Cartier modules

We generalize the Generic Vanishing theorem by Hacon and Patakfalvi in the spirit of Pareschi and Popa. We give several examples illustrating the pathologies appearing in the positive characteristic setting.

preprint2014arXiv

Optimality guarantees for distributed statistical estimation

Large data sets often require performing distributed statistical estimation, with a full data set split across multiple machines and limited communication between machines. To study such scenarios, we define and study some refinements of the classical minimax risk that apply to distributed settings, comparing to the performance of estimators with access to the entire data. Lower bounds on these quantities provide a precise characterization of the minimum amount of communication required to achieve the centralized minimax risk. We study two classes of distributed protocols: one in which machines send messages independently over channels without feedback, and a second allowing for interactive communication, in which a central server broadcasts the messages from a given machine to all other machines. We establish lower bounds for a variety of problems, including location estimation in several families and parameter estimation in different types of regression models. Our results include a novel class of quantitative data-processing inequalities used to characterize the effects of limited communication.

preprint2014arXiv

Pluri-canonical maps of varieties of maximal Albanese dimension in positive characteristic

We show that if X is a nonsingular projective variety of general type over an algebraically closed field k of positive characteristic and X has maximal Albanese dimension and the Albanese map is separable, then |4K_X| induces a birational map.

preprint2014arXiv

Spectral Methods meet EM: A Provably Optimal Algorithm for Crowdsourcing

Crowdsourcing is a popular paradigm for effectively collecting labels at low cost. The Dawid-Skene estimator has been widely used for inferring the true labels from the noisy labels provided by non-expert crowdsourcing workers. However, since the estimator maximizes a non-convex log-likelihood function, it is hard to theoretically justify its performance. In this paper, we propose a two-stage efficient algorithm for multi-class crowd labeling problems. The first stage uses the spectral method to obtain an initial estimate of parameters. Then the second stage refines the estimation by optimizing the objective function of the Dawid-Skene estimator via the EM algorithm. We show that our algorithm achieves the optimal convergence rate up to a logarithmic factor. We conduct extensive experiments on synthetic and real datasets. Experimental results demonstrate that the proposed algorithm is comparable to the most accurate empirical approach, while outperforming several other recently proposed methods.

preprint2013arXiv

Comunication-Efficient Algorithms for Statistical Optimization

We analyze two communication-efficient algorithms for distributed statistical optimization on large-scale data sets. The first algorithm is a standard averaging method that distributes the $N$ data samples evenly to $\nummac$ machines, performs separate minimization on each subset, and then averages the estimates. We provide a sharp analysis of this average mixture algorithm, showing that under a reasonable set of conditions, the combined parameter achieves mean-squared error that decays as $\order(N^{-1}+(N/m)^{-2})$. Whenever $m \le \sqrt{N}$, this guarantee matches the best possible rate achievable by a centralized algorithm having access to all $\totalnumobs$ samples. The second algorithm is a novel method, based on an appropriate form of bootstrap subsampling. Requiring only a single round of communication, it has mean-squared error that decays as $\order(N^{-1} + (N/m)^{-3})$, and so is more robust to the amount of parallelization. In addition, we show that a stochastic gradient-based method attains mean-squared error decaying as $O(N^{-1} + (N/ m)^{-3/2})$, easing computation at the expense of penalties in the rate of convergence. We also provide experimental evaluation of our methods, investigating their performance both on simulated data and on a large-scale regression problem from the internet search domain. In particular, we show that our methods can be used to efficiently solve an advertisement prediction problem from the Chinese SoSo Search Engine, which involves logistic regression with $N \approx 2.4 \times 10^8$ samples and $d \approx 740,000$ covariates.

Institution

Affiliation not imported yet

This author record came from a source that does not expose affiliation metadata. Once the author claims the profile or we enrich the record from another provider, this section will link to the concrete institution.

Source provenance

Where this author record came from

arxivconfidence 95%

external id: arxiv:2605.07536:author:2:yuchen-zhang

Imported May 20, 2026Synced May 21, 2026

arxivconfidence 95%

external id: arxiv:2605.18815:author:2:yuchen-zhang

Imported May 20, 2026Synced May 21, 2026

arxivconfidence 95%

external id: arxiv:2605.11202:author:3:yuchen-zhang

Imported May 20, 2026Synced May 21, 2026

arxivconfidence 95%

external id: arxiv:2604.28082:author:2:yuchen-zhang