Searcharxiv⌕ Search

arXiv subjects

Bhaskar Krishnamachari

Publications and source records attributed to Bhaskar Krishnamachari.

At least 19 recordsLinked to original sources
arXiv2026

Jev as an LLM Safety Guardrail: An Empirical Study of Input Filtering, Selective Output Review, and Over-Refusal

External safety filters can prevent unsafe requests from reaching a large language model (LLM), but may also reject legitimate queries. We present an end-to-end empirical evaluation of Jev, a probabilistic decision model, as a safety guardrail for a Qwen2.5-1.5B-based pipeline. Three configurations are compared on 970 benchmark prompts: an unfiltered LLM, Jev input filtering, and tiered Jev input filtering with selective output review. Across 720 prompts labeled harmful (520 AdvBench and 200 XSTest-unsafe), the fraction recorded as answered decreases from 527/720 (73.2%) without filtering to 17/720 (2.4%) with input filtering. On 250 XSTest-safe prompts, however, non-answer/refusal increases from 3/250 (1.2%) to 27/250 (10.8%). Tiered filtering yields 40/720 (5.6%) harmful-prompt answers and 10/250 (4.0%) benign-prompt refusals, with 79/970 (8.1%) of prompts requiring output review. Joint optimization of the tiered thresholds improves macro-F1 from 0.9351 to 0.9512, while reducing harmful-prompt answers and output-review frequency at the same benign-refusal rate. Matched-budget comparisons reveal modest additional gains over a retuned single gate on the evaluation data, although these gains are not consistently observed on a held-out split.

cs.CR↗
arXiv2026

Schedule Repair for DAG Workflows under Link Disruptions

Schedules for directed acyclic graph (DAG) workflows in networked IoT systems are typically computed assuming a static or generally stable network. In contested and adversarial environments, this assumption is not valid. Links degrade and fail due to mobility, interference, and jamming. We study schedule repair: when a link disruption invalidates part of a schedule, how much of it should be rescheduled? We introduce a spectrum of repair policies that vary in repair scope, how much of the pending schedule each may move: wait out the disruption, reroute data around it, reschedule only the affected tasks locally, or reschedule all pending tasks globally. We evaluate each against an oracle and charge every repair a decision latency proportional to the extent to which it moves. Across 100 workload instances spanning synthetic task graphs, RIoTBench pipelines, and WfCommons scientific workflows, each run at five communication-to-computation ratios (CCRs) and disrupted by processes with deliberately different correlation structure, we find that no single scope wins: rerouting nearly erases isolated failures that cost waiting 30%, global repair comes within 4% of the oracle under jamming blackouts, waiting is favored under memoryless link flapping for larger and communication-heavy workloads (the scheduling analog of route-flap damping), self-healing mobility outages reward patience over reaction, and accounting for repair latency erodes large scopes first. We conclude that the scope of the repair should be adapted to the disruption process and the repair cost, rather than fixed by the scheduler.

cs.DC↗
arXiv2026

Evolving Inspectable O-RAN Slicing xApps with LLMs

Open RAN (O-RAN) slicing xApps must adapt resource allocations to changing channel conditions and traffic demands while meeting service-level agreements (SLAs). Deep reinforcement learning can produce adaptive policies, but their allocation rules remain encoded in neural-network parameters. Our goal is to retain this adaptability while making the controller's decision logic directly inspectable and editable by operators. We use a large language model (LLM) to evolve slicing controllers as compact Python programs whose decision logic remains readable and editable after optimization. The LLM proposes and revises candidates offline, while a calibrated simulator scores them, and the selected decision module runs unchanged in the O-RAN control path. On the NSF POWDER 5G testbed, the evolved controller releases resources from a guaranteed slice whose throughput target becomes unattainable under a sustained channel fade, improving best-effort throughput from 158.2 to 228.6 Mbps, a 44.5% gain over the best static allocation. Since the controllers are readable source code, their behavior can be predicted from their equations, defects can be diagnosed by reading the code, and calibration errors can be corrected with one-line edits, reducing SLA misses from 79.9% to 2.2% in one case and more than doubling fitness in another. In a four-slice trace-driven simulation calibrated to the same testbed, evolutionary search achieves higher average evaluation scores than independent prompting at a matched proposal budget, with mean normalized gains on held-out traces of 16.3% for prompting alone, 32.1% for evolution from scratch, and 51.0% for evolution from a starting program.

eess.SY↗
arXiv2026

Spillover Effects under Network Interference When Neighbours' Treatment Effects Are Heterogeneous

Optimising budget-constrained network interventions requires evaluating not just who is connected, but predicting how strongly individual recipients will propagate the treatment's benefits. Existing models predict spillover from neighbours' treatments and attributes. We argue that spillover also depends on how strongly each neighbour responded to its own treatment. We prove that no model in which neighbours' responsiveness enters separately from their treatments can capture this interaction, and we introduce \textsc{SpilloverNet}, a graph neural network designed to preserve neighbour-level response dynamics. Because a neighbour's response is unobserved, a natural approach is to estimate it from covariates and plug it in. However, we prove that any predictor relying solely on standard network data faces an irreducible error floor set by unobserved personal responsiveness. Empirically, at low heterogeneity the plug-in's correlation with the true response looks acceptable while its spillover error is already 24.5\% against an oracle 13.8\%; as heterogeneity grows its error climbs to 51.9\%, worse than using no responsiveness estimate at all. A per-unit estimate from a direct-response measurement, collected in a small pilot before spillover arrives, escapes this bound and cuts regret against oracle-$τ$ targeting by up to 14 percentage points in a budgeted targeting problem. On two real social graphs, \textsc{SpilloverNet} reaches 7.7--8.4\% error, outperforming standard GNNs as well as specialised causal-representation baselines.

cs.CE↗
arXiv2026

ncsim: A Lightweight Simulator for Networked Edge Computing with Wireless Interference Modeling

Evaluating directed acyclic graph (DAG) task schedulers for wireless edge computing requires jointly modeling compute placement and wireless interference, yet many evaluation setups simplify the wireless coupling. This can lead to rank inversions: a scheduler selected under an interference-free model need not remain best when concurrent wireless traffic is included. We present ncsim, a lightweight discrete-event simulator that combines DAG workflow scheduling with an analytical IEEE 802.11 contention and hidden-interference model in a single Python package. A 108-run factorial experiment on grid networks and DAG workflows finds rank inversions in 6 of 18 scenarios (33.3%). Solo retains single-link MAC overhead, while Full adds concurrent-link wireless effects. Selecting a policy under Solo and executing it under Full incurs a maximum scheduling regret (makespan penalty relative to the best evaluated policy under Full) of 52.3%. Validation includes Bianchi's published numerical and graphical results and fixed-rate ns-3 experiments. The default model closely tracks homogeneous contention. The scheduler results show how wireless coupling can change policy selection within the stated flow model.

cs.DC↗
arXiv2026

A Theory of Information Architecture for Networked Decisions: Freshness, Locality, and Coordination

Networked systems face a tradeoff between the scope of the information behind a decision and its freshness: a broader view of the system supports better coordination, but assembling and communicating it takes time, so it arrives older. We study this tradeoff for a team of agents that repeatedly choose actions to minimize a shared quadratic cost driven by an environment that evolves on its own, unaffected by the agents' actions. An information architecture specifies what each agent observes, from where, and with what delay; we measure an architecture by the cost it loses relative to a decision made with complete, current information. We first compare fresh local observations with a complete but delayed global view. When every component of the environment decorrelates at a common exponential rate, this comparison reduces to a closed-form threshold on the ratio of delay to coherence time. We then study intermediate architectures in which each agent acts on a time-aligned, and therefore older, snapshot of a wider neighborhood. The optimal neighborhood radius occurs where the marginal value of added scope equals the marginal cost of lost freshness. In a canonical spatial model, this radius is set by the spatial-correlation, decision-relevance, and temporal-propagation lengths. Throughout, architecture performance is governed by the predictability of the optimal decision rather than of the raw state.

eess.SY↗
arXiv2026

Scaling Bayesian Bandit Encoding with Shared Learning

A communication system must choose error protection and decoding effort as channel conditions change. A Bayesian bandit encoder (BBE) uses receiver feedback to learn which transmission configuration to select. We study a receiver that decodes by guessing error patterns, using Guessing Random Additive Noise Decoding (GRAND). We extend BBE's selection component to 1,008 code and decoder configurations by learning shared performance patterns offline and updating their weights online. Decoder noise models remain fixed. On a six-configuration training-selected shortlist, sharing reduces accumulated utility loss by 33.5% relative to independent learning. A fixed training-selected configuration matches the shared learner that searches the full catalog. After channel changes, the pruned shared learner first meets a near-optimal selection criterion in 88.5% of events by 2,000 packets, compared with 54.2% for pruned independent learning with the same discounting. The results support combining sharing and pruning for configuration selection, although packet losses remain high for the tested codes under severe noise.

cs.IT↗
arXiv2026

Joint Network-and-Server Congestion in Multi-Source Traffic Allocation: A Convex Formulation and Price-Based Decentralization (Extended Version)

This paper studies an important rate allocation problem that arises in many networked and distributed systems: steady-state traffic rate allocation from multiple sources to multiple service nodes when both (i) the access-path delay on each source-node route is rate-dependent (capacity-constrained) and convex, and (ii) each service node (also capacity-constrained) experiences a load-dependent queueing delay driven by aggregate load from all sources. We show that the resulting flow-weighted end-to-end delay minimization is a convex program, yielding a global system-optimal solution characterized by KKT conditions that equalize total marginal costs (a path marginal access term plus a node congestion price) across all utilized routes. This condition admits a Wardrop-type interpretation: for each source, all utilized options equalize total marginal cost, while any option with strictly larger total marginal cost receives no flow. Building on this structure, we develop a lightweight distributed pricing-based algorithm in which each service node locally computes and broadcasts a scalar congestion price from its observed aggregate load, while each source updates its traffic split by solving a small separable convex allocation problem under the advertised prices. Numerical illustrations demonstrate convergence of the distributed iteration to the centralized optimum and highlight the trade-offs induced by jointly modeling access and service congestion.

cs.DC↗
arXiv2026

Compression-Aware Abstention: Teaching LLMs to Refuse When KV-Compression Masks Remove Answer Evidence

KV-cache compression reduces LLM inference memory by evicting context tokens, but when the evicted tokens contain answer-bearing evidence, the model may hallucinate instead of recognizing that the compressed context is insufficient. We address this failure from a behavioral perspective: to our knowledge, this is the first work to formulate compression-aware abstention as a learning problem, in which a model learns to answer when supporting evidence survives compression and abstain when it does not. We construct supervision from compressor survival masks and tight answer-bearing spans, labeling examples as Confident when evidence survives and Abstain when it is removed. A 10.1M-parameter LoRA adapter trained on ~2.6K MuSiQue 2-hop QA examples reduces base-model hallucinations by 97% under prompt-style truncation while preserving correct answering on evidence-retaining examples. Unlike prompt-only abstention baselines, which over-abstain on many answerable high-retention examples, the trained adapter learns a conditional policy. We also evaluate the method under actual compressed-cache decoding, where multi-compressor training yields a 6-22x relative lift over the unaided base on evidence-retaining examples. Controlled-deletion experiments show that the learned behavior is driven by evidence content rather than input length alone.

cs.CL↗
arXiv2026

Localizing Safety Alignment: MLP Layers and Mid-Network Blocks Encode Refusal Behavior in Large Language Models

Safety alignment in large language models is often treated as a distributed property of the entire network, yet its practical brittleness suggests that refusal behavior may be concentrated in a smaller set of parameters. This work addresses where safety-aligned refusal is encoded by transplanting weights from aligned models into matched unaligned base models at multiple levels of granularity. Using two open-weight model pairs and four safety benchmarks, we conducted experiments to compare the effects of replacing attention weights, MLP weights, contiguous layer regions, and MLP blocks. Across both model families, refusal transfer is dominated by MLP weights: replacing MLP parameters recovers substantially more malicious-prompt refusal than replacing attention parameters, with gains of at least 2.7 times more across benchmarks. Within the MLP stack, refusal-relevant parameters exhibit a consistent mid-network concentration, as the block spanning layers 8-11 is selected first in all six greedy searches over model-dataset pairs. The results also show that the composition of safety-relevant components is non-additive: in five of six greedy trajectories, adding more aligned blocks can reduce refusal performance, and selective block subsets can outperform full MLP transplantation on malicious refusal, benign over-refusal, or both. Finally, greedy orders transferred to OR-Bench vary with the source benchmark used to derive them, indicating a benchmark-dependent precision-coverage trade-off. These results suggest that safety alignment in current LLMs is both localized and interaction-sensitive, offering insight into alignment brittleness and potential avenues for targeted safety interventions.

cs.AI↗
arXiv2026

Closed-Loop Bayesian Bandit Encoder with GRAND Receiver for a Bursty Interference Channel

Interleaving mitigates burst errors but introduces decoding delay and removes temporal error structure that a channel-aware decoder could exploit. We consider packet-level selection between a random linear code and the same code used with cross-codeword interleaving, over a channel with an unknown number of on/off interferers. The receiver uses Guessing Random Additive Noise Decoding (GRAND) with a replaceable noise model and feeds aggregate channel statistics back to a Bayesian estimator at the transmitter. Once the interference amplitudes and timing parameters are estimated, the receiver's noise model is replaced: it computes hidden-Markov-model posterior bit-flip probabilities and uses them to order GRAND queries. A discounted Thompson sampler selects between the two transmission modes using a goodput-minus-latency reward whose distribution is endogenously nonstationary: receiver adaptation, rather than channel change, alters the value of each mode. Across five simulation seeds, the interleaved mode is preferred before channel estimation converges. After the learned decoder is activated, the non-interleaved mode becomes preferable because it achieves lower block error rate without interleaving delay. In the reference configuration, the learned noise model reduces block error rate by approximately one order of magnitude relative to ORBGRAND. Using partial channel estimates before full convergence reduces pre-convergence block error rate by up to $4.5\times$. Adding model-predicted utilities as confidence-weighted pseudo-observations reduces post-transition selection of the inferior arm by approximately $65\%$. Under an idealized airtime conversion at a 100~MHz 5G~NR-like symbol rate, the learning transient corresponds to a few milliseconds of occupied symbol time.

cs.IT↗
arXiv2026

Profiling and Scheduling Complex O-RAN Applications Across the 5G Edge and Cloud

The O-RAN paradigm decomposes intelligent RAN control into pipelines of interdependent AI/ML functions, including traffic prediction, signal quality estimation, and slice scheduling, that must execute across a dispersed continuum of far-edge, near-edge, and cloud resources under heterogeneous latency and bandwidth constraints. Despite the natural expression of these pipelines as Directed Acyclic Graphs (DAGs), no integrated methodology exists to profile their execution costs, map them onto dispersed infrastructure via scheduling heuristics, and validate the resulting placement under 5G cellular conditions. We present O-DAG, an end-to-end framework that closes this gap through four tightly coupled stages: (1) DagProfiler, a new open-source tool that instruments O-RAN Slice Scheduler and extracts per-task instruction counts and per-edge communication volumes; (2) a parameterized three-tier network topology encoding far-edge (DU, RIC), near-edge (edge), and cloud nodes with realistic link bandwidths; (3) an extension of the SAGA scheduling framework and (4) a custom DAG simulation module built on the MintEDGE simulator. We evaluate five scheduling algorithms (HEFT, MCT, MinMin, MaxMin, Duplex) for a slice scheduling application across various configurations spanning 5K--50K UEs, 2--20 cells, and 2--10 network slices. HEFT achieves the lowest makespan in all configurations, but scheduler rankings are workload-dependent. The SAGA--simulation gap serves as a regime diagnostic: negative gaps (up to -1.72%) identify compute-dominated configurations where HEFT overestimates conservatively, while a positive gap (+0.64%) at low slice counts exposes a communication-bound regime where bandwidth contention exceeds the scheduling model's assumptions. All artifacts are released for reproducibility.

cs.DC↗
arXiv2026

Game-Theoretic Framework for Private Data Sharing in Vehicular Networks

We present a novel game-theoretic framework designed to enhance privacy and scalability in decentralized vehicular data collection systems. The proposed hybrid architecture comprises vehicles that supply sensor data, independent servers that process data via secure multiparty computation, a coordinator node that manages data flow, and data consumers that set economic incentives. Crucially, our framework ensures that only the data consumer can access the fully aggregated data, preventing individual raw data exposure and significantly reducing privacy risks. By integrating principles of the Stackelberg competition from game theory, our approach dynamically balances privacy and economic incentives, enabling vehicles to make participation decisions based on perceived privacy risks and incentives. We empirically validate our framework using real-world vehicular location data, quantifying privacy risks by evaluating the accuracy with which a potential adversary can reconstruct a vehicle's path using only a subset of the shared data. This paper details the development and deployment of a data-trading platform within this framework, introducing a practical and privacy-preserving marketplace for profitable vehicle data sharing. Through experiments and simulations, we evaluate the effectiveness of the system in preserving privacy and explore the dynamics that influence vehicle participation. Our findings highlight the robustness of the proposed framework in preserving privacy while supporting an active data market.

cs.NI↗
arXiv2026

TS-Verkle: A TypeScript Native Verkle Library With On-chain Verifier

Blockchain systems face significant scalability challenges due to growing data volumes and increasing transaction demands, necessitating more efficient data structures and verification mechanisms. Verkle trees, a novel data structure combining the efficiency of Merkle trees with the compactness of vector commitments, have gained attention for their potential to optimize blockchain storage and improve scalability. However, their practical implementation, especially at the smart contract level, has remained unexplored. To address these challenges, we present TS-verkle, the first known TypeScript-native implementation of Verkle trees designed for web3 backend compatibility, coupled with a corresponding on-chain verifier written in Solidity. Our work bridges this gap by providing a concrete implementation of Verkle trees and demonstrating their feasibility for on-chain verification. While previous literature suggests Verkle trees should outperform Merkle trees due to their succinct proof size, our empirical evaluation reveals that basic implementations of Verkle trees actually incur higher costs than Merkle trees without advanced optimization techniques. This finding represents a crucial insight for blockchain developers and researchers considering Verkle tree adoption. The paper discusses implementation strategies and performance characteristics while exploring implications for scaling and data availability in decentralized blockchain systems.

cs.DC↗
arXiv2026

Who Restores the Peg? A Mean-Field Game Approach to Model Stablecoin Market Dynamics

USDC and USDT are the dominant stablecoins pegged to \$1 with a total market capitalization of over \$300B and rising. Stablecoins make dollar value globally accessible with secure transfer and settlement. Yet in practice, these stablecoins experience periods of stress and de-pegging from their \$1 target, posing significant systemic risks. The behavior of market participants during these stress events and the collective actions that either restore or break the peg are not well understood. This paper addresses the question: who restores the peg?. We develop a dynamic, agent-based mean-field game framework for fiat-collateralized stablecoins, in which a large population of arbitrageurs and retail traders strategically interact across primary and secondary markets during a de-peg episode. The key advantage of this equilibrium formulation is that it endogenously maps market frictions into a market-clearing price path and implied net order flows, allowing us to attribute peg-reverting pressure by channel and to stress-test when a given infrastructure becomes insufficient for recovery. Using three historical de-peg events, we show that the calibrated equilibrium reproduces observed recovery half-lives and yields an order flow decomposition in which system-wide stress is predominantly stabilized by primary-market arbitrage. Finally, a quantitative sensitivity analysis identifies a non-linear breakdown threshold, beyond which a de-peg becomes markedly slower to reverse.

q-fin.TR↗
arXiv2026

Proactive Market Making and Liquidity Analysis for Everlasting Options in DeFi Ecosystems

Everlasting options, a relatively new class of perpetual financial derivatives, have emerged to tackle the challenges of rolling contracts and liquidity fragmentation in decentralized finance markets. This paper offers an in-depth analysis of markets for everlasting options, modeled using a dynamic proactive market maker. We examine the behavior of funding fees and transaction costs across varying liquidity conditions. Using simulations and modeling, we demonstrate that liquidity providers can aim to achieve a net positive PnL by employing effective hedging strategies, even in challenging environments characterized by low liquidity and high transaction costs. Additionally, we provide insights into the incentives that drive liquidity providers to support the growth of everlasting option markets and highlight the significant benefits these instruments offer to traders as a reliable and efficient financial tool.

q-fin.CP↗
arXiv2026

On the Fluctuations of the Single-Letter $d$-Tilted Sum for Binary Markov Sources

We study the source-side single-letter $d$-tilted sum for a stationary binary Markov chain under Hamming distortion, induced by the single-letter Blahut--Arimoto operating point computed from the stationary marginal $π$. We show that this quantity inherits the same algebraic structure as in the memoryless (i.i.d.) case: the centered sum $J_n(D)-nμ_D$ is exactly an affine function of the chain's occupation count $N_n$, and consequently all centered cumulants are independent of the distortion level $D$. The exact finite-$n$ distribution therefore follows immediately from known results on occupation counts of two-state Markov chains. The genuinely new contributions of this note are (i) a closed-form expression for the finite-$n$ variance that includes the autocorrelation factor due to memory, and (ii) the transfer-matrix representation of the cumulant generating function. The connection, if any, between this source-side quantity and the operational finite-blocklength rate-distortion function remains open.

cs.IT↗
arXiv2026

How to Do Statistical Evaluations in ECE/CS Papers: A Practical Playbook for Defensible Results

Strong experimental papers in electrical and computer engineering and computer science (ECE/CS), especially in systems, networking, and applied machine learning, rest on more than a single impressive number. They rest on a chain of design, measurement, analysis, and validation choices that, taken together, make a result believable. This tutorial is a compact, example-driven guide to that chain for beginning researchers. We organize it as an evaluation workflow: claim, hypothesis, unit of analysis, baseline, regime sweep, uncertainty estimate, validation check, and reporting. Within that workflow we cover the classical statistical foundations (descriptive statistics, the central limit theorem, normal- and $t$-based confidence intervals, Student's $t$-test, ANOVA, chi-squared and Pearson correlation, linear regression) alongside the modern, distribution-free techniques (the bootstrap, Wilcoxon and Mann--Whitney tests, Cliff's delta) that are usually preferred for ECE/CS data. We also discuss factorial design, randomization and blocking, multiple-comparison correction, latency-specific pitfalls, simulation verification and validation, equivalence-style claims, and reproducibility. A running example, a comparison of two job-scheduling algorithms on simulated workloads with truncated heavy-tailed job sizes, threads through the tutorial, with Python snippets the reader can paste and adapt. The paper closes with a pre-submission checklist; companion student-facing material (project-type translation tables, an evaluation-plan worksheet, exercises, and a worked ``bad evaluation autopsy'') is collected in a separate workbook released alongside this paper.

stat.ME↗