Keyboard shortcuts

Press ← or → to navigate between chapters

Press S or / to search in the book

Press ? to show this help

Press Esc to hide this help

research-topic coverage map (authored by agents unless marked 🧑)

scope and evidence

  • snapshot: 7 October 2026 UTC
  • local human wording is evidence of mention
    • talk notes are the human’s notes about speakers
    • they do not independently verify speakers’ claims
  • destinations below identify ownership or a review gap
    • a folder assignment is not proof that a finished study exists
    • this map does not claim exhaustive coverage of every file in the repository
  • sweep examined reading notes, research topic notes, about page, class-note topic matches, and inherited worker handoff
  • a later omission audit inspected 52 additional original inputs
    • selected research-root and cs-root notes plus twelve direct-interest pointers
    • excluded marked agent additions when attributing human interests
    • no confirmed new residual topic found in those inputs
    • unknown-authorship deletion-after-breach question and incomplete RAG traceback methods initially routed to existing owners
    • later manager instruction assigns their uncovered follow-ups to this folder
    • this bounded audit does not establish whole-repository completeness
  • another bounded audit screened 23 previously unchecked files
    • selected soft, government, financial, storage, documentation, and class notes
    • no verified human-topic omission in that input set
    • undeclared authorship prevents treating driver-containment and museum-funding notes as verified human interests
    • museum funding is being investigated as an agent extension of the confirmed museum interest
    • driver fault containment and recovery is an agent extension from an attribution-limited class-note lead
  • practical-note audit screened 55 top-level files across eight folders
    • source files lack authorship declarations; new leads remain unattributed
    • Trail staging design and package environments studied as agent extensions
    • PostgreSQL query-rewriting and TimescaleDB storage-accounting leads submitted for other-group ownership coordination
    • excludes nested files and does not establish whole-repository completeness
    • follow-up search found no eligible nested files in those eight folders after excluding agent instructions
  • six further folders screened: bug notes, macOS, mathematics, exams, USC, and referrals
    • 26 files, including one stylesheet; authorship undeclared
    • mathematics received heading/problem screening rather than complete equation review
      • follow-up maps all eighteen file titles and selected topic passages
      • complex evaluation, validated approximation, computational algebra/topology, and robust geometry were partial gaps in that audit
      • 8 October continuation adds validated complex computation and algebra/geometry
        • focused methods do not cover every mathematical application or routine course topic
      • computational topology now reviews static and incremental persistence as an agent extension
    • numerical reliability studied as an agent extension
    • desktop indexing/recovery studied as an agent extension
    • automated lesson completion maps to learning outcomes
      • scripts establish completion actions, not retained understanding

explicit interests and questions

reading-note mentions grouped by research family

web content, crawling, and standards

  • destination: web infrastructure studies
  • source: human reading notes
    • 🧑 “JavaScript Dead-Code Elimination via HTTP Range Requests (ThorJS), Ayush Pandey, NSL meeting”
    • 🧑 “HTML5, Web Technology class, Marco Papa”
    • 🧑 “Scrapers selectively respect robots.txt directives: evidence from a large-scale empirical study”
    • 🧑 “On YouTube Search API Use in Research, Alexandros Efstratiou”

multilingual accessibility

ML communication and inference

  • destination: LLM inference systems
  • source: human reading notes
    • 🧑 “Collective Communication Algorithms, Arvin Ghavidel, NSL meeting”
    • 🧑 “StreamEP: Straggler-Tolerant MoE Decoding without Communication Barriers, Shaoyu Wang, NSL meeting”
    • 🧑 “I/O Optimizations for Deep Learning on Distributed/Resource-Constrained Environments (Optimus-CC), Yaeyong Song, NSL meeting”
    • 🧑 “Async MoE, Coulson Liang, on behalf of Shaoyu Wang, NSL meeing”

model reasoning and behavior

distributed debugging and mobile testing

hardware observation, WASM, and runtime performance

cloud resource demand and placement

  • destination: cloud scheduling
  • source: human reading notes
    • 🧑 “Granula Resource Demand Heterogeneity, Coulson Liang”
    • 🧑 “The SAP Cloud Infrastructure Dataset: A Reality Check of Scheduling and Placement of VMs in Cloud Computing, Arno Uhlig”
    • 🧑 “Towards Microsecond-Scale vm Core Provisioning Agility on Serverless Platforms, Yibo Yan, NSL meeting”

mobile, wireless, and immersive media

  • destination: wireless models and measurement
    • selected full methods reviewed separately for RTC message parsing, LiVo streaming, and ReVo loss recovery
  • source: human reading notes
    • 🧑 “LiVo: Toward Bandwidth-adaptive Fully-Immersive Volumetric Video Conferencing, Rajrup Ghosh, NSL meeting”
    • 🧑 “Protocol Compliance in Popular RTC Applications, Peiqing Chen”
    • 🧑 “SplatPose: On-Device Outdoor AR Pose Estimation Using Gaussian Splatting, Rajrup Ghosh, NSL meeting”
    • 🧑 “Hybrid Data-Driven and Simulation-Driven Prediction of mmWave Network Performance, Zihao Feng, NSL meeting”

theory, selection, and design

  • destination: theory talks review
    • gaps: exact vote-delegation paper and proof-carrying-state construction unidentified
  • source: human reading notes
    • 🧑 “Defense: Incentivizing Efficient Delegation without Payments, Curtis Bechtel”
    • 🧑 “Lunch in Theory: Near-Optimal Sparsifiers for Stochastic Knapsack and Assignment Problems, Xinyu Liu”
    • 🧑 “Lunch in Theory: How to Measure Differences in Rankings, Yeganeh Alimohammadi”
    • 🧑 “Theory Lunch: Towards Publicly Verifiable Cryptography: Obfuscation, Fully Homomorphic Encryption, and Proof Carrying State, Miryam Huang”
    • 🧑 “Theory Lunch: Proper Learnability and the Role of Unlabeled Data, Julian Asilis”
    • 🧑 “Theory Lunch: An Equivalence Between Fair Division and Wagering Mechanisms, Jens Witkowski”
    • 🧑 “Theory Lunch: Vote Delegation through the Lens of Metric Distortion, Alan Grayson York”
    • 🧑 “Guest Talk: Computational Homogenization for Inverse Design of Surface-based Inflatables, Yingying (Samara) Ren, ISTA”
    • 🧑 “Thesis proposal: Incentivizing Efficient Delegation without Payments, Curtis Bechtel”
    • 🧑 “Scalable -Means Clustering for Large via Seeded Approximate Nearest-Neighbor Search, Jack Spalding-Jamieson”
    • 🧑 “Full Proportional Justified Representation, Jiasen Liu”
    • 🧑 “What uniform symmetric distro can a shallow circuit produce, Kewen Wu”

private computation

  • destination: private-computation methods
    • FedML-HE and blind-annotation full methods reviewed
    • P3V’s complete proof, rerouting, and benchmark configuration remain unchecked because its full PDF was inaccessible
  • source: human reading notes
    • 🧑 “Defense: Efficiency in Privacy-Preserving Computation via Domain Knowledge, Weizhao Jin (advisor: Srivatsan Ravi)”

network optimization and congestion

  • destination: networking and edge systems
  • source: human reading notes
    • 🧑 “Near-Optimal Online Traffic Engineering, Arvin Ghavidel, NSL meeting”
    • 🧑 “MetaRL, Mahdi Alizadeh, NSL meeting”
    • 🧑 “Patchwork: A Traffic Capture and Analysis Platform for Network Experiments on a Federated Testbed, Nishanth Shyamkumar”
    • 🧑 “Congestion Patterns in a Large-scale RDMA Datacenter, Soudeh Ghorbani”

Internet routing and prefix relationships

  • destination: Internet routing and security
  • source: human reading notes
    • 🧑 “A first look into long-lived BGP zombies, Iliana Maria Xygkou”
    • 🧑 “A Framework to Evaluate MPIC Security using Real-World BGP Announcements, Cyrill Krähenbühl”
    • 🧑 “ru-RPKI-ready: the Road Left to Full ROA Adoption, Deepak Gouda”
    • 🧑 “Replication: A Two Decade Review of Policy Atoms - Tracing the Evolution of AS Path Sharing Prefixes, Weili Wu”
    • 🧑 “Learning AS-to-Organization Mappings with Borges, Fabián E. Bustamante”
    • 🧑 “Sibling Prefixes: Identifying Similarities in IPv4 and IPv6 Prefixes, Oliver Gasser”
    • 🧑 “ASINT: Learning AS-to-Organization Mapping from Internet Metadata, Yongzhe Xu”

Internet availability, protocol behavior, and scanners

  • destination: below-web Internet measurement
    • open lead: broadband inequity and scanner attribution need deeper coverage
    • context: networking systems
  • source: human reading notes
    • 🧑 “The Developer, the RFC, and the Middlebox: An HTTP/2 Compliance Story, Mahmoud Attia, Ilies Benhabbour”
    • 🧑 “How I learned to stop worrying and love IPv6: Measuring the Internet Readiness for DNS over IPv6, Anja Feldmann”
    • 🧑 “Tracking Internet Disruptions in Ukraine: Insights from Three Years of Active Full Block Scans, Florian Holzbauer”
    • 🧑 “Have you SYN what I see? Analyzing TCP SYN Payloads in the Wild, Dario Ferrero”
    • 🧑 “Fingerprinting QUIC clients, Seungju Lee”
    • 🧑 “Chunk-fu: Fingerprinting QUIC implementations using fragmented frames, Karthik Nishanth Sengottuvelavan”
    • 🧑 “Passively Inferring Network Availability and Configuration from NTP Pool Clients, Paul Chung”
    • 🧑 “How Do You Know My Name? Investigating The Role of Domain Names for Target Reconnaissance among Web and IPv6 Scanners, Sebastian Kappes”
    • 🧑 “Towards a systematic benchmark framework for evaluating darknet-analysis methodologies, Max Gao”
    • 🧑 “The Potential of Erroneous Outbound Traffic Analysis to Unveil Silent Internal Anomalies, Andrea Sordello”
    • 🧑 “Identifying Disruptive Patterns in Internet Background Radiation, Xie Qiu”
    • 🧑 “IMC 2025 Student Workshop Keynote, Arpit Gupta”
    • 🧑 “Job talk: Expectation vs Reality: How Network Abstractions Impact Internet Security, Paul Pearce”

browser permissions, tracking, and app privacy

  • destination: browser permissions, tracking, and app data collection
  • source: human reading notes
    • 🧑 “A Permissions Odyssey: A Systematic Study of Browser Permissions on Modern Websites, Alberto Fernandez-de-Retana”
    • 🧑 “An In-Depth Investigation of Data Collection in LLM App Ecosystems, Yuhao Wu”
    • 🧑 “FP-Inconsistent: Measurement and Analysis of Fingerprint Inconsistencies in Evasive Bot Traffic, Hari Venugopalan”
    • 🧑 “CookieGuard: Characterizing and Isolating the First-Party Cookie Jar, Zubair Shafiq”
    • 🧑 “Canvassing the Fingerprinters: Characterizing Canvas Fingerprinting Use Across the Web, Elisa Luo”
    • 🧑 “Where in the World Are My Trackers? Mapping Web Tracking Flow Across Diverse Geographic Regions, Robert Ricci”

browser GPU performance

peer-to-peer storage and blockchain infrastructure

spam and creator consent

  • destination: spam campaigns and content removal and theft
  • source: human reading notes
    • 🧑 “Do Spammers Dream of Electric Sheep? Characterizing the Prevalence of LLM-Generated Malicious Emails, Wei Hao”
    • 🧑 “Somesite I Used To Crawl: Awareness, Agency and Efficacy in Protecting Content Creators From AI Crawlers, Elisa Luo”

health and anatomical simulation

non-Rust language infrastructure

scientific writing, proposals, and organizational decisions

economics, politics, and personal media notes

attention and studying

recovery of the earlier 18-topic sweep

remaining broad interests that the 18-topic list did not explicitly handle

coverage test

8 October completion and publication audit

  • local paused-worker audit found research gaps separately from publication absence
    • misinformation interventions replaces the earlier unfinished sketch
    • social-media campaigns now has a dedicated review
    • StreamEP and Async MoE were routed to serving without an actual exact-topic treatment
      • asynchronous MoE review now treats the notes and closest primary methods
        • StreamEP manuscript access remains incomplete; its authenticated artifact is separately reviewed
    • cholesterol/diet/heart-disease attribution comes only from the earlier agent’s topic list
  • the existing AI-agent, systems-ML, and web-trust studies now have transferred completion/publication ownership
    • the Claude tasks remain paused
    • local existence alone did not establish publication
  • source review corrected material inherited proposal errors
    • TBIK already demonstrates matching outputs/probabilities across tested tensor-parallel sizes
    • one GPU cannot run a four-device tensor-parallel experiment
    • a passing test cannot establish unreachability
    • unused-code maintenance cost has direct prior studies
  • cross-topic consultation completed at verified Extra High
    • advice and paper leads were assessed separately from scientific evidence
  • completeness limit
    • source discovery and reading remain bounded and stated per page
    • no claim of exhaustive coverage of every repository file or every field
    • no proposed experiment was run and no proposal novelty is certified

Last edited: