BZPEERReading, trust and collaboration for research
Menu

Discover

GraphFollow connections around a paper, topic or saved view.

Workspaces

HomePick up where your research flow left off.InboxHandle requests, replies and notifications from one place.CollectionsCurate reading lists, evidence sets and shareable dossiers.ResearchSee your private provenance graph, trail and imports.CollaborationMove from discovery into focused discussion rooms.PublishDraft, preview and publish full-text research articles.

Network

ExpertsFind specialists worth following or asking for help.CommunitiesJoin research circles organized around shared work.PartnersSee external organizations active around the same topics.

Opportunities

ListingsBrowse roles, calls and collaborations with clearer fit signals.

Account

Log inReturn to your reading and collaboration workspace.Create accountStart saving work, following people and joining teams.
Log inCreate account

Researcher profile

Yuchen Zhao

Yuchen Zhao contributes to research discovery and scholarly infrastructure.

ResearcherAffiliation not importedOpen to collaborate
cond-mat.softMachine Learningcond-mat.stat-mechcond-mat.dis-nncond-mat.mtrl-sciCryptography and Securityeess.SPNetworking and Internet Architecturenlin.AOPerformancephysics.ins-detphysics.opticsRoboticsSocial and Information Networks

Trust snapshot

Quick read

Trust 21 - EmergingVerification L1Unclaimed author
9works
0followers
14topics
4close collaborators

Actions

Decide how to stay connected

Follow researcher0

Identity and collaboration

How to connect with this researcher

Claiming links this public author record to a researcher profile and unlocks direct collaboration workflows.

Log in to claim

Direct collaboration

Open a focused conversation when the fit is right

Claim this author entity first to unlock direct invitations.

Research graph

See the researcher in context

Open full explorer

Inspect adjacent work, topics, institutions and collaborators without jumping out to a separate graph page.

Building this graph slice

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

Published work

9 published item(s)

preprint2026arXiv

Synecdoche: Efficient and Accurate In-Network Traffic Classification via Direct Packet Sequential Pattern Matching

Traffic classification on programmable data plane holds great promise for line-rate processing, with methods evolving from per-packet to flow-level analysis for higher accuracy. However, a trade-off between accuracy and efficiency persists. Statistical feature-based methods align with hardware constraints but often exhibit limited accuracy, while online deep learning methods using packet sequential features achieve superior accuracy but require substantial computational resources. This paper presents Synecdoche, the first traffic classification framework that successfully deploys packet sequential features on a programmable data plane via pattern matching, achieving both high accuracy and efficiency. Our key insight is that discriminative information concentrates in short sub-sequences--termed Key Segments--that serve as compact traffic features for efficient data plane matching. Synecdoche employs an "offline discovery, online matching" paradigm: deep learning models automatically discover Key Segment patterns offline, which are then compiled into optimized table entries for direct data plane matching. Extensive experiments demonstrate Synecdoche's superior accuracy, improving F1-scores by up to 26.4% against statistical methods and 18.3% against online deep learning methods, while reducing latency by 13.0% and achieving 79.2% reduction in SRAM usage. The source code of Synecdoche is publicly available to facilitate reproducibility and further research.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

A photon counting reconstructive spectrometer combining metasurfaces and superconducting nanowire single-photon detectors

Faint light spectroscopy has many important applications such as fluorescence spectroscopy, lidar and astronomical observations. However, long measurement time limit its application on real-time measurement. In this work, a photon counting reconstructive spectrometer combining metasurfaces and superconducting nanowire single photon detectors (SNSPDs) was proposed. A prototype device was fabricated on a silicon on isolator (SOI) substrate, and its performance was characterized. Experiment results show that this device support spectral reconstruction of mono-color lights with a resolution of 2 nm in the wavelength region of 1500 nm ~ 1600 nm. The detection efficiency of this device is 1.4% ~ 3.2% in this wavelength region. The measurement time required by this photon counting reconstructive spectrometer was also investigated experimentally, showing its potential to be applied in the scenarios requiring real-time measurement.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Multimodal Federated Learning on IoT Data

Federated learning is proposed as an alternative to centralized machine learning since its client-server structure provides better privacy protection and scalability in real-world applications. In many applications, such as smart homes with Internet-of-Things (IoT) devices, local data on clients are generated from different modalities such as sensory, visual, and audio data. Existing federated learning systems only work on local data from a single modality, which limits the scalability of the systems. In this paper, we propose a multimodal and semi-supervised federated learning framework that trains autoencoders to extract shared or correlated representations from different local data modalities on clients. In addition, we propose a multimodal FedAvg algorithm to aggregate local autoencoders trained on different data modalities. We use the learned global autoencoder for a downstream classification task with the help of auxiliary labelled data on the server. We empirically evaluate our framework on different modalities including sensory data, depth camera videos, and RGB camera videos. Our experimental results demonstrate that introducing data from multiple modalities into federated learning can improve its classification performance. In addition, we can use labelled data from only one modality for supervised learning on the server and apply the learned model to testing data from other modalities to achieve decent F1 scores (e.g., with the best performance being higher than 60%), especially when combining contributions from both unimodal clients and multimodal clients.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Towards Battery-Free Machine Learning and Inference in Underwater Environments

This paper is motivated by a simple question: Can we design and build battery-free devices capable of machine learning and inference in underwater environments? An affirmative answer to this question would have significant implications for a new generation of underwater sensing and monitoring applications for environmental monitoring, scientific exploration, and climate/weather prediction. To answer this question, we explore the feasibility of bridging advances from the past decade in two fields: battery-free networking and low-power machine learning. Our exploration demonstrates that it is indeed possible to enable battery-free inference in underwater environments. We designed a device that can harvest energy from underwater sound, power up an ultra-low-power microcontroller and on-board sensor, perform local inference on sensed measurements using a lightweight Deep Neural Network, and communicate the inference result via backscatter to a receiver. We tested our prototype in an emulated marine bioacoustics application, demonstrating the potential to recognize underwater animal sounds without batteries. Through this exploration, we highlight the challenges and opportunities for making underwater battery-free inference and machine learning ubiquitous.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Ultra-stable shear jammed granular material

Dry granular materials such as sand, gravel, pills, or agricultural grains, can become rigid when compressed or sheared. At low density, one can distort the shape of a container of granular material without encountering any resistance. Under isotropic compression, the material will reach a certain {\it jamming} density and then resist further compression. {\em Shear jamming} occurs when resistance to shear emerges in a system at a density lower than the jamming density, and the elastic properties of such states have important implications for industrial and geophysical processes. We report on experimental observations of changes in the mechanical properties of a shear-jammed granular material subjected to small-amplitude, quasi-static cyclic shear. We study a layer of plastic discs confined to a shear cell, using photoelasticimetry to measure all inter-particle vector forces. For sufficiently small cyclic shear amplitudes and large enough initial shear, the material evolves to an unexpected "ultra-stable" state in which all the particle positions and inter-particle contact forces remain unchanged after each complete shear cycle for thousands of cycles. The stress response of these states to small imposed shear is nearly elastic, in contrast to the original shear jammed state.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

A Library for Constraint Consistent Learning

This paper introduces the first, open source software library for Constraint Consistent Learning (CCL). It implements a family of data-driven methods that are capable of (i) learning state-independent and -dependent constraints, (ii) decomposing the behaviour of redundant systems into task- and null-space parts, and (iii) uncovering the underlying null space control policy. It is a tool to analyse and decompose many everyday tasks, such as wiping, reaching and drawing. The library also includes several tutorials that demonstrate its use with both simulated and real world data in a systematic way. This paper documents the implementation of the library, tutorials and associated helper methods. The software is made freely available to the community, to enable code reuse and allow users to gain in-depth experience in statistical learning in this area.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

Yielding, Rigidity, and Tensile Stress in Sheared Columns of Hexapod Granules

Granular packings of non-convex or elongated particles can form free-standing structures like walls or arches. For some particle shapes, such as staples, the rigidity arises from interlocking of pairs of particles, but the origins of rigidity for non-interlocking particles remains unclear. We report on experiments and numerical simulations of sheared columns of "hexapods," particles consisting of three mutually orthogonal sphero-cylinders whose centers coincide. We vary the length-to-diameter aspect ratio, $α$, of the sphero-cylinders and subject the packings to quasistatic direct shear. For small $α$, we observe a finite yield stress. For large $α$, however, the column becomes rigid when sheared, supporting stresses that increase sharply with increasing strain. Analysis of X-ray micro-computed tomography (Micro-CT) data collected during the shear reveals that the stiffening is associated with a tilted, oblate cluster of hexapods near the nominal shear plane in which particle deformation and average contact number both increase. Simulation results show that the particles are collectively under tension along one direction even though they do not interlock pairwise. These tensions comes from contact forces carrying large torques, and they are perpendicular to the compressive stresses in the packing. They counteract the tendency to dilate, thus stabilize the particle cluster.

0 citations0 reviewsNo code linkedOpen access
preprint2019arXiv

Using Metrics Suites to Improve the Measurement of Privacy in Graphs

Social graphs are widely used in research (e.g., epidemiology) and business (e.g., recommender systems). However, sharing these graphs poses privacy risks because they contain sensitive information about individuals. Graph anonymization techniques aim to protect individual users in a graph, while graph de-anonymization aims to re-identify users. The effectiveness of anonymization and de-anonymization algorithms is usually evaluated with privacy metrics. However, it is unclear how strong existing privacy metrics are when they are used in graph privacy. In this paper, we study 26 privacy metrics for graph anonymization and de-anonymization and evaluate their strength in terms of three criteria: monotonicity indicates whether the metric indicates lower privacy for stronger adversaries; for within-scenario comparisons, evenness indicates whether metric values are spread evenly; and for between-scenario comparisons, shared value range indicates whether metrics use a consistent value range across scenarios. Our extensive experiments indicate that no single metric fulfills all three criteria perfectly. We therefore use methods from multi-criteria decision analysis to aggregate multiple metrics in a metrics suite, and we show that these metrics suites improve monotonicity compared to the best individual metric. This important result enables more monotonic, and thus more accurate, evaluations of new graph anonymization and de-anonymization algorithms.

0 citations0 reviewsNo code linkedOpen access
preprint2015arXiv

Packings of 3D stars: Stability and structure

We describe a series of experiments involving the creation of cylindrical packings of star-shaped particles, and an exploration of the stability of these packings. The stars cover a broad range of arm sizes and frictional properties. We carried out three different kinds of experiments, all of which involve columns that are prepared by raining star particles one-by-one into hollow cylinders. As an additional part of the protocol, we sometimes vibrated the column before removing the confining cylinder. We rate stability in terms of r, the ratio of the mass of particles that fall off a pile when it collapsed, to the total particle mass. The first experiment involved the intrinsic stability of the pile when the confining cylinder was removed. The second kind of experiment involved adding a uniform load to the top of the column, and then determining the collapse properties. A third experiment involved testing stability to tipping of the piles. We find a stability diagram relating the pile height, h, vs. pile diameter, delta, where the stable and unstable regimes are separated by a boundary that is roughly a power-law in h vs. delta with an exponent that is less than one. Increasing friction and vibration both tend to stabilize piles, while increasing particle size can destabilize the system under certain conditions.

0 citations0 reviewsNo code linkedOpen access