Pith. sign in

REVIEW 3 major objections 4 minor 35 references

To make an internet of AI agents discoverable, the paper proposes a telemetry protocol (DOVIS) and a ranking algorithm (AgentRank-UC) that fuses usage and competence through two coupled PageRank fixed points, with formal guarantees of uniqu

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

T0 review · deepseek-v4-flash

2026-08-05 05:44 UTC pith:5JCWVI3O

load-bearing objection Useful architecture proposal and a clean contraction result, but the monotonicity proof has a real gap and the Sybil-resistance claim is much narrower than the abstract suggests. the 3 major comments →

arxiv 2509.04979 v1 pith:5JCWVI3O submitted 2025-09-05 cs.AI

Internet 3.0: Architecture for a Web-of-Agents with it's Algorithm for Ranking Agents

Rajesh Tembarai Krishnamachari , Srividya Rajesh This is my paper
classification cs.AI
keywords Agentic Webagent rankingusage-competence fusionPageRank fixed pointtelemetry protocolSybil resistancecold-start fairnesscompetence-aware discovery
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper argues that the emerging Web of Agents—where autonomous AI agents, not humans, are the primary users of the internet—will be unusable unless there is a principled way to choose which agent to call for a task. It claims that such a ranking can be built from minimal, privacy-preserving telemetry: callers publish only aggregate counts of calls, successes, quality, latency, cost, and risk, and an indexer combines these into two page-rank-like scores, one for usage and one for competence. The central result is that this fused ranking, AgentRank-UC, is unique, converges quickly, never punishes improved performance, gives newcomers a nonzero baseline visibility, and cannot be overwhelmed by Sybil usage-pumping. If correct, the paper supplies the missing discovery substrate for an agent economy: a way to rank agents by proven recent behavior rather than declared capability or raw popularity. The authors support the claim with formal theorems and simulations using archetypal agents, shocks, and colluding cliques.

Core claim

The paper proposes that the Agentic Web—a network where AI agents delegate tasks to one another—needs a global, competence-aware ranking, and claims that this ranking is achievable with a five-layer protocol (DOVIS: Discovery, Orchestration, Verification, Incentives, Semantics) plus an algorithm called AgentRank-UC. AgentRank-UC builds two row-stochastic kernels from decayed telemetry: a usage kernel P capturing who calls whom, and a competence kernel Q capturing how well those calls turn out. The usage rank x and competence rank y are the fixed points of x = αP^T x + (1−α)v and y = βQ^T y + (1−β)w, combined into r = normalize(x^p ⊙ y^{1−p}). The paper proves that these fixed points exist, a

What carries the argument

Two coupled PageRank-style operators over the agent graph: x = αP^T x + (1−α)v (usage rank) and y = βQ^T y + (1−β)w (competence rank), with teleportation priors v and w, fused by the geometric mean r = normalize(x^p ⊙ y^{1−p}). The usage and competence kernels P and Q are built from exponentially decayed aggregates (calls, successes, quality, latency, cost, risk), and the teleportation terms are what make the ranking well-posed: they are contraction mappings with unique fixed points, and they enforce a strict positivity floor for every agent, which is the mechanism behind both cold-start fairness and the Sybil non-amplification bound.

Load-bearing premise

The ranking's guarantees rest on callers reporting honest aggregates: if callers can inflate successes or suppress failures without being caught by signatures, callee acknowledgments, and the 1–5% audit rate, the competence kernel Q is corrupted and the theorems no longer describe real agent behavior.

What would settle it

Run a live or simulated marketplace with DOVIS telemetry for dozens of epochs, and independently measure each agent's true competence with held-out test calls; if a colluding clique that passes the audit checks still captures top ranks through inflated n_success or suppressed n_failure, the competence kernel is corrupted and the monotonicity and Sybil guarantees fail in practice.

Watch this falsifier. Get emailed when new claim-graph text bears on it.

Share X Bluesky LinkedIn Reddit HN

If this is right

  • If AgentRank-UC is correct, an open ecosystem of agents can be ranked for discovery using only minimal aggregate telemetry, without exposing raw prompts, responses, or user data.
  • The fixed-point formulation gives a computationally concrete path: linear-time construction of sufficient statistics and O(log 1/ε) power iterations per ranking, each dominated by a sparse matrix-vector product.
  • Monotonicity ensures that improving success rate, quality, or latency never lowers an agent's rank, removing the perverse incentive to hide improvements.
  • Cold-start fairness follows from strictly positive priors: newcomers always retain a minimum visibility share, with a tunable floor via the teleportation parameters.
  • The Sybil non-amplification bound shows that a clique of colluding agents can inflate usage only up to a hard ceiling, and cannot capture the fused rank unless it also improves genuine competence.
  • Simulations indicate the ranking can be tuned by the balance parameter p and the half-life H, giving operators an interpretable trade-off between popularity, competence, responsiveness, and stability.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The paper's formal guarantees treat the kernels P and Q as fixed inputs, but in deployment the ranking feeds back into caller selection, so the usage kernel co-evolves with the rank itself; a natural extension is to analyze the coupled dynamical system where selection probabilities depend on r.
  • The Sybil bound leans on competence share y_S remaining below 1; if colluding agents can also inflate competence on easy self-tests, the resistance weakens, suggesting a testable boundary where competence manipulation, not usage pumping, becomes the limiting attack.
  • The balance parameter p could be made task- or sparsity-dependent rather than a global constant; the paper lists this as future work, and it is a concrete way to let discovery favor competence in sparse niches and usage in mature ones.
  • The DOVIS incentive layer is described qualitatively; a quantitative mechanism-design question—what reporting cost, audit probability, and penalty make honest reporting a Nash equilibrium—is left open and is directly testable from the paper's parameters.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

3 major / 4 minor

Summary. The paper proposes DOVIS, a five-layer protocol (Discovery, Orchestration, Verification, Incentives, Semantics) for collecting minimal, signed aggregate telemetry in an open ecosystem of AI agents, and AgentRank-UC, a ranking algorithm that combines a usage PageRank fixed point x = αPᵀx + (1−α)v and a competence fixed point y = βQᵀy + (1−β)w into a fused score r = normalize(xᵖ ⊙ y^{1−p}). The manuscript presents theorems asserting existence, uniqueness, linear convergence, continuity, monotonicity, cold-start fairness, perturbation stability, and Sybil non-amplification, together with simulations on synthetic worlds with archetypal agents, shocks, and adversarial cliques. The central claim is that AgentRank-UC is a well-posed, competence-aware ranking substrate for the Agentic Web when built on DOVIS telemetry.

Significance. If the theoretical guarantees held in full, the paper would provide a useful concrete starting point for competence-aware agent discovery, combining a minimal telemetry schema with a clean two-kernel PageRank-style algorithm. The contraction fixed-point results (Theorems 4.1 and 4.13) are standard but correctly presented, and the explicit Sybil mass bound in Theorem 4.17 is a nice illustration of how teleportation limits usage-only amplification. The DOVIS protocol is described at a level that could inform prototype implementations. However, two load-bearing issues limit the significance: the monotonicity proof in Theorem 4.7 has a genuine gap, and the Sybil-resistance and outcome-monotonicity guarantees are conditional on competence telemetry being truthful, an assumption the paper itself identifies as the most serious threat but never models formally. The simulations are a useful sanity check but do not close those gaps.

major comments (3)
  1. [§4.4, Theorem 4.7 (proof Step 2)] The proof asserts without argument that "Row-normalization preserves monotonicity in the target column j: increasing C_ij raises the relative share of mass directed to j," and then concludes that the fixed point y* is monotone in each entry of Q. This is not a standard fact and is generally false: increasing C_ij increases Q_ij but decreases Q_ik for k≠j, so the operator T_Q does not increase entrywise in Q. A perturbation calculation for y = βQᵀy + (1−β)w shows the sign of dy_j depends on differences of Green's function entries (I − βQᵀ)^{-1}; there is no reason it is sign-definite. The monotonicity theorem is therefore not established as stated. Since "monotone in outcomes" is a headline guarantee, this gap is load-bearing.
  2. [§2.3, §5.9, Theorem 4.17] Theorem 4.17's Sybil bound (2) and the Sybil-resistance conclusions in §5.9 assume "collusion does not improve competence (i.e., Q is unchanged on S)". But §2.3 explicitly calls success inflation and failure suppression the "most serious threat," and the Exp-5 simulation only pumps usage while Sybils honestly report mediocre competence. No theorem bounds the distortion of Q under strategic misreporting, and no experiment simulates success inflation. Thus the advertised Sybil resistance and the related monotonicity guarantee are guarantees about reported outcomes, not about real competence in the open adversarial setting that motivates DOVIS. The verification/incentive layer needs a formal model, or the claims must be restated as conditional on honest or audit-bounded telemetry.
  3. [§5.9, Exp-5] Even within the honest-telemetry setting, the Sybil experiment is too narrow to support the general robustness claim: it considers a single configuration of clique size, link density, and utility penalties, reports only the final epoch (plus a trajectory plot), and gives no error bars or multiple seeds. More importantly, the experiment does not vary the audit rate, penalty strength, or prior weighting that DOVIS's verification layer would introduce, so it does not test the protocol-level defenses that the paper argues are essential. This is a secondary issue relative to the unmodeled success-inflation threat, but it should be addressed in a revision.
minor comments (4)
  1. [Title] The title reads "it's Algorithm"; should be "its Algorithm."
  2. [§5.2 and Exp-1/2/3/4] The experiments repeatedly refer to "archetypes in Section 4.2," but the archetypes are defined in Section 5.2 (World Model); Section 4 is the theory section. Please fix the cross-reference.
  3. [Figure 1] Figure 1 (radar chart of archetypes) is referenced but no image appears in the provided text. Ensure the figure is actually included.
  4. [§3.3, Step 3] The utility weights θ1...θ5 are introduced without guidance on calibration or default values. If they are free parameters, the paper should say so explicitly and discuss sensitivity; if they are part of the protocol, their semantics should be specified.

Circularity Check

0 steps flagged

No significant circularity: the paper's guarantees are explicit consequences of its stated definitions, with the Sybil bound openly conditional on unchanged competence.

full rationale

The paper's formal results (Theorems 4.1, 4.4, 4.7, 4.10, 4.13, and 4.17) are conditional derivations from the explicitly defined AgentRank-UC equations, not circular predictions. Existence/uniqueness follows from the contraction property of the PageRank-style operators; monotonicity follows from stated isotone utilities and monotone aggregation; cold-start positivity follows from strictly positive priors; and the Sybil bound explicitly assumes that collusion does not change Q on the Sybil set, an assumption the paper states rather than hides. DOVIS's verification threats are described in Section 2.3 as 'the most serious threat,' and the proofs do not attempt to cover success inflation mathematically, which is a limitation or correctness risk, not a circularity. The simulations are synthetic and do not benchmark against external ground truth, but the p-sweep is explicitly framed as an interpolation that coincides with competence-only at p=0 and usage-only at p=1 by the definition of the geometric fusion, and Experiment 1 labels the naive success rate an 'oracle' only in a caller-independent world where that rate converges to true competence. No fitted parameter is renamed as a prediction, no load-bearing claim rests on a self-citation, and no equation is reduced to its own input. The derivation chain is self-contained; the main caveats are honesty/verification assumptions and lack of adversarial modeling of report inflation, which are outside the mathematical circularity criteria.

Axiom & Free-Parameter Ledger

7 free parameters · 8 axioms · 2 invented entities

The central ranking formula contains several designer-set hyperparameters (α, β, p, λ, utility weights, smoothing counts, priors). None are estimated from real data, and the protocol and telemetry schema are proposed without a deployed implementation. The mathematical guarantees are relative to these chosen parameters and to the assumption that reported telemetry is honest or effectively audited.

free parameters (7)
  • teleport weights α and β = not specified
    Designer-chosen weights in the usage and competence fixed-point equations; they control contraction rates and all theoretical bounds.
  • balance parameter p = swept over [0,1] in Exp-2
    Controls the geometric fusion of usage and competence; no principled selection method is given.
  • recency decay half-life H (λ) = H in {4,8,16} epochs in simulations
    Determines how quickly rankings forget old telemetry; a free design parameter traded against stability.
  • utility weights θ1...θ5 = not reported
    Weights balancing success, quality, latency, cost, and risk in the competence utility; no calibration method or values are provided.
  • Beta smoothing pseudo-counts α0, β0 = not reported
    Hyperparameters in the smoothed success estimator; affect small-sample behavior but are never specified.
  • routing parameters ρ, γ, exploration ε, temperature τ = not reported
    Caller selection-rule parameters in simulations; without values the simulation results are not fully determined.
  • priors v and w = uniform in most simulations
    Baseline distributions for cold-start fairness; theoretical results only require strict positivity, leaving the choice open.
axioms (8)
  • standard math Banach fixed-point theorem and contraction mapping theory
    Used in Theorem 4.1 to prove existence, uniqueness, and linear convergence.
  • standard math Row-stochastic matrix and Markov chain theory
    Background for PageRank-style fixed points and perturbation bounds (Theorems 4.1 and 4.13).
  • standard math Hölder's inequality and convexity of the geometric mean
    Used in the Sybil non-amplification bound of Theorem 4.17.
  • domain assumption Exponential decay recency kernel is an adequate model of relevance
    Section 3.2 assumes ω(t) = e^{-λ(T-t)} captures the importance of recent performance without empirical validation.
  • domain assumption Caller-side aggregate statistics are sufficient for ranking
    Section 2.1 asserts that counts, successes, and sums of quality, latency, cost, and risk are all that is needed; no argument rules out missing relevant signals.
  • domain assumption Telemetry is honest or verifiable via signatures, audits, and incentives
    Sections 2.3 and 2.4 assume the verification and incentive layers deter misreporting sufficiently for the competence kernel to remain trustworthy.
  • ad hoc to paper Competence is reducible to a linear-log utility in success, quality, latency, cost, and risk
    The utility form in Step 3 of Section 3.3 is chosen without justification; the weights are uncalibrated.
  • ad hoc to paper Geometric fusion x^p ⊙ y^{1-p} is the appropriate combination of usage and competence
    Section 3.3 selects the geometric mean to penalize imbalance, but no derivation or comparative evaluation justifies this specific operator over alternatives.
invented entities (2)
  • OAT-Lite telemetry schema no independent evidence
    purpose: A minimal aggregate reporting format for callers to publish usage and performance signals.
    Proposed standard; no deployed implementation, interoperability test, or external adoption evidence is provided.
  • DOVIS five-layer protocol no independent evidence
    purpose: Coordination substrate for discovery, orchestration, verification, incentives, and semantics in agent ecosystems.
    Design proposal validated only by synthetic simulation and theoretical arguments, not by a working system.

pith-pipeline@v1.4.0-alltime-deepseek-medium · 24985 in / 19037 out tokens · 197462 ms · 2026-08-05T05:44:06.895416+00:00 · methodology

0 comments
Cite this review

Pith. "Pith review of Internet 3.0: Architecture for a Web-of-Agents with it's Algorithm for Ranking Agents." pith.science (2026). https://pith.science/paper/5JCWVI3O

@misc{pith2026250904979,
  author       = {Pith},
  title        = {Pith review of: Internet 3.0: Architecture for a Web-of-Agents with it's Algorithm for Ranking Agents},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/5JCWVI3O}},
  note         = {Machine review of arXiv:2509.04979}
}
Share X Bluesky LinkedIn Reddit HN
read the original abstract

AI agents -- powered by reasoning-capable large language models (LLMs) and integrated with tools, data, and web search -- are poised to transform the internet into a \emph{Web of Agents}: a machine-native ecosystem where autonomous agents interact, collaborate, and execute tasks at scale. Realizing this vision requires \emph{Agent Ranking} -- selecting agents not only by declared capabilities but by proven, recent performance. Unlike Web~1.0's PageRank, a global, transparent network of agent interactions does not exist; usage signals are fragmented and private, making ranking infeasible without coordination. We propose \textbf{DOVIS}, a five-layer operational protocol (\emph{Discovery, Orchestration, Verification, Incentives, Semantics}) that enables the collection of minimal, privacy-preserving aggregates of usage and performance across the ecosystem. On this substrate, we implement \textbf{AgentRank-UC}, a dynamic, trust-aware algorithm that combines \emph{usage} (selection frequency) and \emph{competence} (outcome quality, cost, safety, latency) into a unified ranking. We present simulation results and theoretical guarantees on convergence, robustness, and Sybil resistance, demonstrating the viability of coordinated protocols and performance-aware ranking in enabling a scalable, trustworthy Agentic Web.

Figures

Figures reproduced from arXiv: 2509.04979 by Rajesh Tembarai Krishnamachari, Srividya Rajesh.

Figure 1
Figure 1. Figure 1: Desirability radar for agent archetypes across four dimensions: Competence, Speed (inverse [PITH_FULL_IMAGE:figures/full_fig_p021_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Experiment 1 bar charts side-by-side. between them as p varies. This construction avoids feedback-loop artifacts in dynamic simulations and provides a clear visualization of how the merge parameter governs the trade-off. We instantiate the same agent world as in Exp-1, with n = 100 agents spanning d = 3 task types and archetypes (Popular-but-Mediocre, Niche-but-Excellent, Balanced-Strong, Cheap-but-Risky, … view at source ↗
Figure 3
Figure 3. Figure 3: Experiment 2b results: UC interpolates between competence-only (p [PITH_FULL_IMAGE:figures/full_fig_p024_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Experiment 3 rank trajectories: degraded PbM agent vs. improved NbE agent under varying [PITH_FULL_IMAGE:figures/full_fig_p025_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Experiment 4 diagnostic tests: (a) monotonicity and (b) cold-start behavior. [PITH_FULL_IMAGE:figures/full_fig_p026_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Experiment 5: SybilMass comparison. (a) Sybil mass by method at final epoch. (b) Sybil [PITH_FULL_IMAGE:figures/full_fig_p027_6.png] view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

35 extracted references · 24 canonical work pages

  1. [1]

    Abul Ehtesham, Aditi Singh, Gaurav Kumar Gupta, and Saket Kumar. A survey of agent interop- erability protocols: Model context protocol (mcp), agent communication protocol (acp), agent- to-agent protocol (a2a), and agent network protocol (anp).arXiv preprint arXiv:2505.02279, 2025

    Pith/arXiv arXiv 2025

  2. [2]

    Agent network protocol technical white paper.arXiv preprint arXiv:2508.00007, July 2025

    Agent Network Protocol (ANP) Open Source Technology Community, Gaowei Chang, Eidan Lin, Chengxuan Yuan, Rizhao Cai, Binbin Chen, Xuan Xie, and Yin Zhang. Agent network protocol technical white paper.arXiv preprint arXiv:2508.00007, July 2025. Posted on arXiv: July 18, 2025

    Pith/arXiv arXiv 2025

  3. [3]

    O’Brien, Carrie J

    Joon Sung Park, Joseph C. O’Brien, Carrie J. Cai, Meredith R. Morris, Percy Liang, and Michael S. Bernstein. Generative agents: Interactive simulacra of human behavior. InarXiv preprint arXiv:2304.03442, 2023

    Pith/arXiv arXiv 2023

  4. [4]

    CAMEL: Communicative agents for “mind” exploration of large language model society

    Guohao Li, Hasan Abed Al Kader Hammoud, Hadi Itani, Dmitry Khizbullin, and Bernard Ghanem. CAMEL: Communicative agents for “mind” exploration of large language model society. InarXiv preprint arXiv:2303.17760, 2023

    Pith/arXiv arXiv 2023

  5. [5]

    AutoGen: Enabling next-gen llm applications via multi-agent conversation

    Qian Wu, Gagan Bansal, Jiaming Zhang, and et al. AutoGen: Enabling next-gen llm applications via multi-agent conversation. InarXiv preprint arXiv:2308.08155, 2023

    Pith/arXiv arXiv 2023

  6. [6]

    AgentBench: Evaluating LLMs as agents

    Xiaoyu Liu, Haotian Yu, Haoran Zhang, and et al. AgentBench: Evaluating LLMs as agents. In International Conference on Learning Representations (ICLR), 2024

    work page 2024

  7. [7]

    Reflexion: Language agents with verbal reinforcement learning

    Nathaniel Shinn, Francesco Cassano, Ethan Berman, and et al. Reflexion: Language agents with verbal reinforcement learning. InNeural Information Processing Systems (NeurIPS), 2023

    work page 2023

  8. [8]

    V oyager: An open-ended embodied agent with large language models

    Guanzhi Wang, Yilun Xie, Yuke Jiang, Ajay Mandlekar, Chenguang Xiao, Yuke Zhu, Linxi Fan, and Anima Anandkumar. V oyager: An open-ended embodied agent with large language models. InarXiv preprint arXiv:2305.16291, 2023

    Pith/arXiv arXiv 2023

  9. [9]

    Agentic web: Weaving the next web with ai agents.arXiv preprint arXiv:2507.21206, 2025

    Yingxuan Yang, Mulei Ma, Yuxuan Huang, Huacan Chai, Chenyu Gong, Haoran Geng, Yuanjian Zhou, Ying Wen, Meng Fang, Muhao Chen, Shangding Gu, Ming Jin, Costas Spanos, Yang Yang, Pieter Abbeel, Dawn Song, Weinan Zhang, and Jun Wang. Agentic web: Weaving the next web with ai agents.arXiv preprint arXiv:2507.21206, 2025

    Pith/arXiv arXiv 2025

  10. [10]

    The anatomy of a large-scale hypertextual web search engine

    Sergey Brin and Lawrence Page. The anatomy of a large-scale hypertextual web search engine. InProceedings of the 7th International World Wide Web Conference (WWW7), 1998

    work page 1998

  11. [11]

    Kleinberg

    Jon M. Kleinberg. Authoritative sources in a hyperlinked environment.Journal of the ACM, 46(5):604–632, 1999

    work page 1999

  12. [12]

    Haveliwala

    Taher H. Haveliwala. Topic-sensitive PageRank. InProceedings of the 11th International World Wide Web Conference (WWW), 2002

    work page 2002

  13. [13]

    Haveliwala, Sepandar D

    Taher H. Haveliwala, Sepandar D. Kamvar, and Glen Jeh. An analytical comparison of approaches to personalizing pagerank. Technical report, Stanford University, 2003. 29

    work page 2003

  14. [14]

    Langville and Carl D

    Amy N. Langville and Carl D. Meyer.Google’s PageRank and Beyond: The Science of Search Engine Rankings. Princeton University Press, 2006

    work page 2006

  15. [15]

    David F. Gleich. Pagerank beyond the web.SIAM Review, 57(3):321–363, 2015

    2015

  16. [16]

    Saltzer, David P

    Jerome H. Saltzer, David P. Reed, and David D. Clark. End-to-end arguments in system design. ACM Transactions on Computer Systems, 2(4):277–288, 1984

    work page 1984

  17. [17]

    David D. Clark. The design philosophy of the DARPA internet protocols. InProceedings of the ACM SIGCOMM Symposium, 1988

    work page 1988

  18. [18]

    Information technology — open systems interconnection: Basic reference model

    ISO/IEC. Information technology — open systems interconnection: Basic reference model. ISO/IEC 7498-1:1994, 1994

    work page 1994

  19. [19]

    QUIC: A UDP-based multiplexed and secure transport

    Jana Iyengar and Martin Thomson. QUIC: A UDP-based multiplexed and secure transport. RFC 9000, IETF, 2021

    work page 2021

  20. [20]

    Sigelman, Luiz André Barroso, Mike Burrows, and et al

    Benjamin H. Sigelman, Luiz André Barroso, Mike Burrows, and et al. Dapper, a large-scale distributed systems tracing infrastructure. InGoogle Technical Report, 2010

    work page 2010

  21. [21]

    The OpenTelemetry specification.https://opentelemetry.io/, 2023

    CNCF. The OpenTelemetry specification.https://opentelemetry.io/, 2023

    work page 2023

  22. [22]

    Practical secure aggregation for privacy-preserving machine learning

    Keith Bonawitz, Vladimir Ivanov, Ben Kreuter, and et al. Practical secure aggregation for privacy-preserving machine learning. InProceedings of the ACM Conference on Computer and Communications Security (CCS), 2017

    work page 2017

  23. [23]

    Faster private set intersection based on OT extension

    Benny Pinkas, Thomas Schneider, and Michael Zohner. Faster private set intersection based on OT extension. InProceedings of the 23rd USENIX Security Symposium, 2014

    work page 2014

  24. [24]

    Fast and private computation of cardinality of set intersection and union

    Emiliano De Cristofaro, Paolo Gasti, and Gene Tsudik. Fast and private computation of cardinality of set intersection and union. InProceedings of the 17th International Conference on Information Security, 2012

    work page 2012

  25. [25]

    Intel SGX explained

    Victor Costan and Srinivas Devadas. Intel SGX explained. IACR Cryptology ePrint Archive 2016/086, 2016

    work page 2016

  26. [26]

    Brendan McMahan, Eider Moore, Daniel Ramage, Seth Hampson, and Blaise Aguera y Arcas

    H. Brendan McMahan, Eider Moore, Daniel Ramage, Seth Hampson, and Blaise Aguera y Arcas. Communication-efficient learning of deep networks from decentralized data. InProceedings of the 20th International Conference on Artificial Intelligence and Statistics (AISTATS), 2017

    work page 2017

  27. [27]

    The algorithmic foundations of differential privacy.Founda- tions and Trends in Theoretical Computer Science, 9(3-4):211–407, 2014

    Cynthia Dwork and Aaron Roth. The algorithmic foundations of differential privacy.Founda- tions and Trends in Theoretical Computer Science, 9(3-4):211–407, 2014

    2014

  28. [28]

    Deep learning with differential privacy

    Martin Abadi, Andy Chu, Ian Goodfellow, and et al. Deep learning with differential privacy. In Proceedings of the ACM Conference on Computer and Communications Security (CCS), 2016

    work page 2016

  29. [29]

    Combating web spam with trustrank

    Zoltán Gyöngyi, Hector Garcia-Molina, and Jan Pedersen. Combating web spam with trustrank. InProceedings of the 30th International Conference on Very Large Data Bases (VLDB), 2004

    work page 2004

  30. [30]

    Levin, Yuval Peres, and Elizabeth L

    David A. Levin, Yuval Peres, and Elizabeth L. Wilmer.Markov Chains and Mixing Times. American Mathematical Society, 2nd edition, 2017

    2017

  31. [31]

    Horn and Charles R

    Roger A. Horn and Charles R. Johnson.Matrix Analysis. Cambridge University Press, 2nd edition, 2013

    2013

  32. [32]

    Springer, 2nd edition, 2006

    Eugene Seneta.Non-negative Matrices and Markov Chains. Springer, 2nd edition, 2006

    work page 2006

  33. [33]

    Gibbons, Michael Kaminsky, and Abraham D

    Haifeng Yu, Phillip B. Gibbons, Michael Kaminsky, and Abraham D. Flaxman. SybilGuard: Defending against Sybil attacks via social networks. InProceedings of ACM SIGCOMM, 2006

    work page 2006

  34. [34]

    Gibbons, Michael Kaminsky, and Feng Xiao

    Haifeng Yu, Phillip B. Gibbons, Michael Kaminsky, and Feng Xiao. SybilLimit: A near-optimal social network defense against Sybil attacks. InProceedings of IEEE Symposium on Security and Privacy (Oakland), 2008

    work page 2008

  35. [35]

    Kamvar, Mario T

    Sepandar D. Kamvar, Mario T. Schlosser, and Hector Garcia-Molina. EigenTrust: Reputation management in peer-to-peer networks. InProceedings of the 12th International World Wide Web Conference (WWW), 2003. 30

    work page 2003