arXiv subjects

Kumar Sharad

Publications and source records attributed to Kumar Sharad.

3 recordsLinked to original sources

arXiv2026

Observability for Delegated Execution in Agentic AI Systems

Delegation-scoped execution is not identifiable from standard observables: audit logs and execution traces can be identical under multiple incompatible delegation assignments. This gap is especially acute in LLM-based agentic systems, where agents dynamically select tools, vary execution sequences across runs for the same instruction, and spawn cooperating sub-agents. These dynamics fragment and interleave traces, making delegation-scoped reconstruction from causal structure alone structurally underdetermined. Although individual actions are authorized and logged, existing audit, tracing, and security schemas lack the semantics to reconstruct what actions occurred under a given delegation across heterogeneous systems. We focus on delegation-scoped attribution and access/share footprint reconstruction, not intent inference or reasoning reconstruction. We present an agent-aware observability substrate consisting of a lightweight gateway and a common information model that binds delegation context at execution time. This enables reliable cross-tool delegation-scoped reconstruction and direct forensic queries without heuristic time-window correlation.

cs.CR↗

arXiv2018

On the Security of Randomized Defenses Against Adversarial Samples

Deep Learning has been shown to be particularly vulnerable to adversarial samples. To combat adversarial strategies, numerous defensive techniques have been proposed. Among these, a promising approach is to use randomness in order to make the classification process unpredictable and presumably harder for the adversary to control. In this paper, we study the effectiveness of randomized defenses against adversarial samples. To this end, we categorize existing state-of-the-art adversarial strategies into three attacker models of increasing strength, namely blackbox, graybox, and whitebox (a.k.a.~adaptive) attackers. We also devise a lightweight randomization strategy for image classification based on feature squeezing, that consists of pre-processing the classifier input by embedding randomness within each feature, before applying feature squeezing. We evaluate the proposed defense and compare it to other randomized techniques in the literature via thorough experiments. Our results indeed show that careful integration of randomness can be effective against both graybox and blackbox attacks without significantly degrading the accuracy of the underlying classifier. However, our experimental results offer strong evidence that in the present form such randomization techniques cannot deter a whitebox adversary that has access to all classifier parameters and has full knowledge of the defense. Our work thoroughly and empirically analyzes the impact of randomization techniques against all classes of adversarial strategies.

cs.CR↗

arXiv2014

An Automated Social Graph De-anonymization Technique

We present a generic and automated approach to re-identifying nodes in anonymized social networks which enables novel anonymization techniques to be quickly evaluated. It uses machine learning (decision forests) to matching pairs of nodes in disparate anonymized sub-graphs. The technique uncovers artefacts and invariants of any black-box anonymization scheme from a small set of examples. Despite a high degree of automation, classification succeeds with significant true positive rates even when small false positive rates are sought. Our evaluation uses publicly available real world datasets to study the performance of our approach against real-world anonymization strategies, namely the schemes used to protect datasets of The Data for Development (D4D) Challenge. We show that the technique is effective even when only small numbers of samples are used for training. Further, since it detects weaknesses in the black-box anonymization scheme it can re-identify nodes in one social network when trained on another.

cs.CR↗