arXiv subjects

Ziao Wang

Publications and source records attributed to Ziao Wang.

At least 19 recordsLinked to original sources

arXiv2026

Achieving Almost Exact Recovery in Almost Quadratic Time: Rank-Based Graph Matching via Local Tree Correlation Tests

This paper studies graph matching under the correlated $\text{Erd\H{o}s-R\'{e}nyi}$ (ER) graph pair model. This model first samples an $\mathrm{ER}(n,\frac{\lambda}{ns})$ base graph, whose edges are then independently subsampled twice with probability $s$ to produce two correlated $\mathrm{ER}(n,\frac{\lambda}{n})$ graphs. We propose a graph matching algorithm that has $n^{2+o(1)}$ time complexity and achieves almost exact recovery with high probability under the assumptions $\lambda=(\log n)^{\alpha+o(1)}$ for some $\alpha\in(0,1)$ and $s\in(\sqrt{C_{\mathrm{Otter}}},1]$, where $C_{\mathrm{Otter}}\approx 0.338$ is Otter's tree-counting constant. This is the first algorithm with almost quadratic time complexity in this regime of $\lambda$, while the best known result in this regime is the chandelier-counting algorithm with time complexity $O(n^{c(s)})$, where $c(s)\rightarrow \infty$ as $s$ approaches $\sqrt{C_\mathrm{Otter}}$ from above. The proposed algorithm is based on local tree correlation tests. It uses a rank-based algorithm to match the vertex pairs instead of threshold-based rules in the literature. This avoids the need of computing an explicit threshold, which is computationally difficult to obtain. To prove the almost exact recovery result, we establish a new analysis of tree correlation tests in the diverging-degree regime, where both the mean degree and the tree depth grow with $n$. Based on this new result, we establish the existence of a threshold for a threshold-based graph matching algorithm via local tree correlation tests. Finally, we couple the performance of the rank-based algorithm with the threshold-based algorithm to show almost exact recovery.

cs.DS↗

arXiv2026

Diffusion-Network Alignment: An Efficient Algorithm and Explicit Probability Bounds

This paper studies a variation of the classic network alignment problem, named diffusion-network alignment. The goal is to align the vertices of a rooted diffusion tree to the vertices of a network, where the diffusion tree could be from a communication trace or contact tracing, and the network could be an online or offline social network. Different from the classic network alignment where both networks are fully observed, this model captures the information asymmetry of two networks. To solve this problem, this paper presents an efficient algorithm based on tree correlation tests to extract alignment information from local neighborhoods. We analyze the performance of the algorithm in the sparse graph regime and show that with high probability, all matched pairs are correct. Furthermore, for each vertex on the diffusion tree, this paper establishes an explicit lower bound on the probability that the vertex is correctly matched. These lower bounds are depth-dependent and increase as vertices get closer to the root.

cs.DS↗

arXiv2026

GRACE: Gradient-aligned Reasoning Data Curation for Efficient Post-training

Existing reasoning data curation pipelines score whole samples, treating every intermediate step as equally valuable. In reality, steps within a trace contribute very unevenly, and selecting reasoning data well requires assessing them individually. We present GRACE, a gradient-aligned curation method that views each reasoning trace as a sequence of optimization events and scores every step by two complementary signals: its alignment with the answer-oriented gradient direction, and its consistency with the preceding reasoning trajectory. Step-level scores are aggregated into a sample-level value for subset selection, using only the model's internal optimization signals and no external reward models or step annotations. To make this scalable, GRACE introduces a representation-level gradient proxy that estimates step-level alignment from token-level upstream signals in a single forward pass. Post-training Qwen3-VL-2B-Instruct on MMathCoT-1M, GRACE reaches 108.8% of the full-data performance with 20% of the data and retains 100.2% with only 5%, with subsets that transfer effectively across model backbones.

cs.AI↗

arXiv2026

TRON: Trainable, architecture-reconfigurable random optical neural networks

Deep learning has triggered explosive growth in the demand for specialized hardware processors, thus motivating the development of scalable and reconfigurable computing substrates. Optical processors offer a fundamentally different computing paradigm, combining massive parallelism and ultrahigh bandwidth with the potential for substantial energy savings. However, progress has been constrained by the absence of scalable and reconfigurable architectures that can implement a broad class of network architectures. Here, we introduce TRON, a scalable and trainable optoelectronic deep optical neural network that exploits a multi-scattering medium and a DMD as a learnable, high-dimensional dense optical matrix multiplier, processing with fixed and tunable optical operations. We perform in-situ optimization of the optical parameters involved in the scattering process, together with automated neural architecture search (NAS) and optimization directly on optics. The experimental results demonstrate that in-situ NAS is essential to discover architectures that adapt to both the task and hardware constraints, establishing a viable path towards large-scale optical processors for next-generation machine learning and data-intensive computing.

physics.optics↗

arXiv2026

Multi-Agent SAC Enabled Beamforming Design for Joint Secret Key Generation and Data Transmission

Physical layer key generation (PLKG) has emerged as a promising solution for achieving highly secured and low-latency key distribution, offering information-theoretic security that is inherently resilient to quantum attacks. However, simultaneously ensuring a high data transmission rate and a high secret key generation rate under eavesdropping attacks remains a major challenge. In time-division duplex (TDD) systems with multiple antennas, we derive closed-form expressions for both rates by modeling the legitimate channel as a time-correlated autoregressive (AR) process. This formulation leads to a highly nonconvex and time-coupled optimization problem, rendering traditional optimization methods ineffective. To address this issue, we propose a multi-agent soft actor-critic (SAC) framework equipped with a long short-term memory (LSTM) adversary prediction module to cope with the partial observability of the eavesdropper's mode. Simulation results demonstrate that the proposed approach achieves superior performance compared with other benchmark algorithms, while effectively balancing the trade-off between secret key generation rate and data transmission rate. The results also confirm the robustness of the proposed framework against intelligent eavesdropping and partial observation uncertainty.

cs.IT↗

arXiv2026

Training deep physical neural networks with local physical information bottleneck

Deep learning has revolutionized modern society but faces growing energy and latency constraints. Deep physical neural networks (PNNs) are interconnected computing systems that directly exploit analog dynamics for energy-efficient, ultrafast AI execution. Realizing this potential, however, requires universal training methods tailored to physical intricacies. Here, we present the Physical Information Bottleneck (PIB), a general and efficient framework that integrates information theory and local learning, enabling deep PNNs to learn under arbitrary physical dynamics. By allocating matrix-based information bottlenecks to each unit, we demonstrate supervised, unsupervised, and reinforcement learning across electronic memristive chips and optical computing platforms. PIB also adapts to severe hardware faults and allows for parallel training via geographically distributed resources. Bypassing auxiliary digital models and contrastive measurements, PIB recasts PNN training as an intrinsic, scalable information-theoretic process compatible with diverse physical substrates.

cs.LG↗

arXiv2026

Uncovering Intrinsic Capabilities: A Paradigm for Data Curation in Vision-Language Models

Large vision-language models (VLMs) achieve strong benchmark performance, but controlling their behavior through instruction tuning remains difficult. Reducing the budget of instruction tuning dataset often causes regressions, as heuristic strategies treat models as black boxes and overlook the latent capabilities that govern learning. We introduce Capability-Attributed Data Curation (CADC), a framework that shifts curation from task-specific heuristics to intrinsic capability analysis. CADC discovers intrinsic capabilities in an unsupervised manner from gradient-based learning trajectories, attributes training data to these capabilities via influence estimation, and curates capability-aware curricula through balanced selection and staged sequencing. This transforms black-box instruction tuning into a controllable, capability-driven process. With as little as 5% of the original data, CADC surpasses full-data training on multimodal benchmarks. These results validate intrinsic capabilities as the fundamental building blocks of model learning and establish CADC as a principle paradigm for instruction data curation.

cs.CV↗

arXiv2025

Test-Time Steering for Lossless Text Compression via Weighted Product of Experts

Lossless compression techniques are crucial in an era of rapidly growing data. Traditional universal compressors like gzip offer low computational overhead, high speed, and broad applicability across data distributions. However, they often lead to worse compression rates than modern neural compressors, which leverage large-scale training data to model data distributions more effectively. Despite their advantages, neural compressors struggle to generalize to unseen data. To address this limitation, we propose a novel framework that performs Test-Time Steering via a Weighted Product of Experts (wPoE). At inference, our method adaptively combines a universal compression model with a pretrained neural language model, ensuring the compression rate is at least as good as that of the best individual model. Extensive experiments demonstrate that our approach improves the performance of text compression without requiring fine-tuning. Furthermore, it seamlessly integrates with any autoregressive language model, providing a practical solution for enhancing text compression across diverse data distributions.

cs.CL↗

arXiv2025

Nexus: An Omni-Perceptive And -Interactive Model for Language, Audio, And Vision

This work proposes an industry-level omni-modal large language model (LLM) pipeline that integrates auditory, visual, and linguistic modalities to overcome challenges such as limited tri-modal datasets, high computational costs, and complex feature alignments. Our pipeline consists of three main components: First, a modular framework enabling flexible configuration of various encoder-LLM-decoder architectures. Second, a lightweight training strategy that pre-trains audio-language alignment on the state-of-the-art vision-language model Qwen2.5-VL, thus avoiding the costly pre-training of vision-specific modalities. Third, an audio synthesis pipeline that generates high-quality audio-text data from diverse real-world scenarios, supporting applications such as Automatic Speech Recognition and Speech-to-Speech chat. To this end, we introduce an industry-level omni-modal LLM, Nexus. Extensive experiments validate the efficacy of our pipeline, yielding the following key findings:(1) In the visual understanding task, Nexus exhibits superior performance compared with its backbone model - Qwen2.5-VL-7B, validating the efficiency of our training strategy. (2) Within the English Spoken Question-Answering task, the model achieves better accuracy than the same-period competitor (i.e, MiniCPM-o2.6-7B) in the LLaMA Q. benchmark. (3) In our real-world ASR testset, Nexus achieves outstanding performance, indicating its robustness in real scenarios. (4) In the Speech-to-Text Translation task, our model outperforms Qwen2-Audio-Instruct-7B. (5) In the Text-to-Speech task, based on pretrained vocoder (e.g., Fishspeech1.4 or CosyVoice2.0), Nexus is comparable to its backbone vocoder on Seed-TTS benchmark. (6) An in-depth analysis of tri-modal alignment reveals that incorporating the audio modality enhances representational alignment between vision and language.

cs.MM↗

arXiv2025

Optical Computing with Spectrally Multiplexed Features in Complex Media

Artificial intelligence (AI) has rapidly evolved into a critical technology; however, electrical hardware struggles to keep pace with the exponential growth of AI models. Free space optical hardware provides alternative approaches for large-scale optical processing, and in-memory computing, with applications across diverse machine learning tasks. Here, we explore the use of broadband light scattering in free-space optical components, specifically complex media, which generate uncorrelated optical features at each wavelength. By treating individual wavelengths as independent predictors, we demonstrate improved classification accuracy through in-silico majority voting, along with the ability to estimate uncertainty without requiring access to the model's probability outputs. We further demonstrate that linearly combining multiwavelength features, akin to spectral shaping, enables us to tune output features with improved performance on classification tasks, potentially eliminating the need for multiple digital post-processing steps. These findings illustrate the spectral multiplexing or broadband advantage for free-space optical computing.

physics.optics↗

arXiv2025

Two-dimensional magnetic tunnel p-n junctions for low-power electronics

For decades, semiconductors and their heterostructures have underpinned both fundamental and applied research across all areas of electronics. Two-dimensional, 2D (atomically thin) semiconductors have now the potential to push further the miniaturization of electronic components, enabling the development of more efficient electronics. Here, we report on a giant anomalous zero-bias spin voltage in magnetic tunnel junctions based on 2D materials. The generation, manipulation and detection of electron spin across a nanometer-thick magnetic tunnel junction do not require any applied bias. It is achieved by exploiting high-quality ferromagnetic/semiconductor interfaces and the asymmetric diffusion of spin-up/spin-down electrons across a semiconductor p-n junction. The large spin-voltage signal exceeds 30,000% and is far greater than the highest magnetoresistance signals reported to date. Our findings reveal unexplored opportunities to transform and amplify spin information for low-power electronics.

cond-mat.mes-hall↗

arXiv2024

Noisy Computing of the Threshold Function

Let $\mathsf{TH}_k$ denote the $k$-out-of-$n$ threshold function: given $n$ input Boolean variables, the output is $1$ if and only if at least $k$ of the inputs are $1$. We consider the problem of computing the $\mathsf{TH}_k$ function using noisy readings of the Boolean variables, where each reading is incorrect with some fixed and known probability $p \in (0,1/2)$. As our main result, we show that it is sufficient to use $(1+o(1)) \frac{n\log \frac{m}δ}{D_{\mathsf{KL}}(p \| 1-p)}$ queries in expectation to compute the $\mathsf{TH}_k$ function with a vanishing error probability $δ= o(1)$, where $m\triangleq \min\{k,n-k+1\}$ and $D_{\mathsf{KL}}(p \| 1-p)$ denotes the Kullback-Leibler divergence between $\mathsf{Bern}(p)$ and $\mathsf{Bern}(1-p)$ distributions. Conversely, we show that any algorithm achieving an error probability of $δ= o(1)$ necessitates at least $(1-o(1))\frac{(n-m)\log\frac{m}δ}{D_{\mathsf{KL}}(p \| 1-p)}$ queries in expectation. The upper and lower bounds are tight when $m=o(n)$, and are within a multiplicative factor of $\frac{n}{n-m}$ when $m=Θ(n)$. In particular, when $k=n/2$, the $\mathsf{TH}_k$ function corresponds to the $\mathsf{MAJORITY}$ function, in which case the upper and lower bounds are tight up to a multiplicative factor of two. Compared to previous work, our result tightens the dependence on $p$ in both the upper and lower bounds.

cs.DS↗

arXiv2024

Surprising Resilience of Science During a Global Pandemic: A Large-Scale Descriptive Analysis

The COVID-19 pandemic profoundly impacted people globally, yet its effect on scientists and research institutions has yet to be fully examined. To address this knowledge gap, we use a newly available bibliographic dataset covering tens of millions of papers and authors to investigate changes in research activity and collaboration during this period. Employing statistical methods, we analyze the pandemic's disruptions on the participation, productivity, and collaborations of researchers at the top 1,000 institutions worldwide based on historical productivity, taking into account variables such as geography, researcher seniority and gender, and field of study. Our findings reveal an unexpected trend: research activity and output significantly increased in the early stages of the pandemic, indicating a surprising resilience in the scientific community. However, by the end of 2022, there was a notable reversion to historical trends in research participation and productivity. This reversion suggests that the initial spike in research activity was a short-lived disruption rather than a permanent shift. As such, monitoring scientific outputs in 2023 and beyond becomes crucial. There may be a delayed negative effect of the pandemic on research, given the long time horizon for many research fields and the temporary closure of wet labs. Further analysis is needed to fully comprehend the factors that underpin the resilience of scientific innovation in the face of global crises. Our study provides an initial comprehensive exploration up to the end of 2022, offering valuable insights into how the scientific community has adapted and responded over the course of the pandemic.

cs.SI↗

arXiv2024

Streamlined optical training of large-scale modern deep learning architectures with direct feedback alignment

Modern deep learning relies nearly exclusively on dedicated electronic hardware accelerators. Photonic approaches, with low consumption and high operation speed, are increasingly considered for inference but, to date, remain mostly limited to relatively basic tasks. Simultaneously, the problem of training deep and complex neural networks, overwhelmingly performed through backpropagation, remains a significant limitation to the size and, consequently, the performance of current architectures and a major compute and energy bottleneck. Here, we experimentally implement a versatile and scalable training algorithm, called direct feedback alignment, on a hybrid electronic-photonic platform. An optical processing unit performs large-scale random matrix multiplications, which is the central operation of this algorithm, at speeds up to 1500 TeraOPS under 30 Watts of power. We perform optical training of modern deep learning architectures, including Transformers, with more than 1B parameters, and obtain good performances on language, vision, and diffusion-based generative tasks. We study the scaling of the training time, and demonstrate a potential advantage of our hybrid opto-electronic approach for ultra-deep and wide neural networks, thus opening a promising route to sustain the exponential growth of modern artificial intelligence beyond traditional von Neumann approaches.

cs.ET↗

arXiv2024

Noisy Sorting Capacity

Sorting is the task of ordering $n$ elements using pairwise comparisons. It is well known that $m=Θ(n\log n)$ comparisons are both necessary and sufficient when the outcomes of the comparisons are observed with no noise. In this paper, we study the sorting problem when each comparison is incorrect with some fixed yet unknown probability $p$. Unlike the common approach in the literature which aims to minimize the number of pairwise comparisons $m$ to achieve a given desired error probability, we consider randomized algorithms with expected number of queries $\textsf{E}[M]$ and aim at characterizing the maximal sorting rate $\frac{n\log n}{\textsf{E}[M]}$ such that the ordering of the elements can be estimated with a vanishing error probability asymptotically. The maximal rate is referred to as the noisy sorting capacity. In this work, we derive upper and lower bounds on the noisy sorting capacity. The two lower bounds -- one for fixed-length algorithms and one for variable-length algorithms -- are established by combining the insertion sort algorithm with the well-known Burnashev--Zigangirov algorithm for channel coding with feedback. Compared with existing methods, the proposed algorithms are universal in the sense that they do not require the knowledge of $p$, while maintaining a strictly positive sorting rate. Moreover, we derive a general upper bound on the noisy sorting capacity, along with an upper bound on the maximal rate that can be achieved by sorting algorithms that are based on insertion sort.

cs.IT↗

arXiv2024

Efficient Algorithms for Attributed Graph Alignment with Vanishing Edge Correlation

Graph alignment refers to the task of finding the vertex correspondence between two correlated graphs of $n$ vertices. Extensive study has been done on polynomial-time algorithms for the graph alignment problem under the Erdős-Rényi graph pair model, where the two graphs are Erdős-Rényi graphs with edge probability $q_\mathrm{u}$, correlated under certain vertex correspondence. To achieve exact recovery of the correspondence, all existing algorithms at least require the edge correlation coefficient $ρ_\mathrm{u}$ between the two graphs to be \emph{non-vanishing} as $n\rightarrow\infty$. Moreover, it is conjectured that no polynomial-time algorithm can achieve exact recovery under vanishing edge correlation $ρ_\mathrm{u}<1/\mathrm{polylog}(n)$. In this paper, we show that with a vanishing amount of additional \emph{attribute information}, exact recovery is polynomial-time feasible under \emph{vanishing} edge correlation $ρ_\mathrm{u} \ge n^{-Θ(1)}$. We identify a \emph{local} tree structure, which incorporates one layer of user information and one layer of attribute information, and apply the subgraph counting technique to such structures. A polynomial-time algorithm is proposed that recovers the vertex correspondence for most of the vertices, and then refines the output to achieve exact recovery. The consideration of attribute information is motivated by real-world applications like LinkedIn and Twitter, where user attributes like birthplace and education background can aid alignment.

cs.DS↗

arXiv2024

Universal Graph Compression: Stochastic Block Models

Motivated by the prevalent data science applications of processing large-scale graph data such as social networks and biological networks, this paper investigates lossless compression of data in the form of a labeled graph. Particularly, we consider a widely used random graph model, stochastic block model (SBM), which captures the clustering effects in social networks. An information-theoretic universal compression framework is applied, in which one aims to design a single compressor that achieves the asymptotically optimal compression rate, for every SBM distribution, without knowing the parameters of the SBM. Such a graph compressor is proposed in this paper, which universally achieves the optimal compression rate with polynomial time complexity for a wide class of SBMs. Existing universal compression techniques are developed mostly for stationary ergodic one-dimensional sequences. However, the adjacency matrix of SBM has complex two-dimensional correlations. The challenge is alleviated through a carefully designed transform that converts two-dimensional correlated data into almost i.i.d. submatrices. The sequence of submatrices is then compressed by a Krichevsky--Trofimov compressor, whose length analysis is generalized to identically distributed but arbitrarily correlated sequences. In four benchmark graph datasets, the compressed files from competing algorithms take 2.4 to 27 times the space needed by the proposed scheme.

cs.IT↗

arXiv2024

On the Feasible Region of Efficient Algorithms for Attributed Graph Alignment

Graph alignment aims at finding the vertex correspondence between two correlated graphs, a task that frequently occurs in graph mining applications such as social network analysis. Attributed graph alignment is a variant of graph alignment, in which publicly available side information or attributes are exploited to assist graph alignment. Existing studies on attributed graph alignment focus on either theoretical performance without computational constraints or empirical performance of efficient algorithms. This motivates us to investigate efficient algorithms with theoretical performance guarantee. In this paper, we propose two polynomial-time algorithms that exactly recover the vertex correspondence with high probability. The feasible region of the proposed algorithms is near optimal compared to the information-theoretic limits. When specialized to the seeded graph alignment problem under the seeded Erdős--Rényi graph pair model, the proposed algorithms extends the best known feasible region for exact alignment by polynomial-time algorithms.

cs.DS↗