Searcharxiv⌕ Search

arXiv subjects

Mayank Gulati

Publications and source records attributed to Mayank Gulati.

4 recordsLinked to original sources

Auditing the Judge: Internal Signals for LLM Reasoning Assurance

Language-model systems use self-judgment to decide which answers to release, review, or recompute. Because intervention is costly, an assurance score must rank errors well at the available budget, not merely over the full ranking. We compare the judge's continuous verdict-token score with a supervised correctness probe trained on attention-head activations from the same prompt. Exact identities link fixed-budget error recall to residual risk and show that AUROC averages recall over all budgets. Across three open 7-12B models on GSM8K and MATH, the probe finds 1.3-22.5 additional percentage points of errors at 5% review in all six model-dataset settings (median 7.4 points). At 10% routing, probe-guided allocation improves accuracy over equal-cost random allocation in both GSM8K experiments, but beats judge-guided routing only for the weaker judge. Cross-dataset transfer lowers probe AUROC for every model in both directions relative to target-domain training. Because dataset, weight precision, and generation budget change together, this measures a joint protocol shift rather than a dataset-only effect. Internal signals can improve supervised, in-domain budget allocation, but are not domain-general confidence estimators.

cs.LG↗

ALIGN-FL: Architecture-independent Learning through Invariant Generative component sharing in Federated Learning

We present ALIGN-FL, a novel approach to distributed learning that addresses the challenge of learning from highly disjoint data distributions through selective sharing of generative components. Instead of exchanging full model parameters, our framework enables privacy-preserving learning by transferring only generative capabilities across clients, while the server performs global training using synthetic samples. Through complementary privacy mechanisms: DP-SGD with adaptive clipping and Lipschitz regularized VAE decoders and a stateful architecture supporting heterogeneous clients, we experimentally validate our approach on MNIST and Fashion-MNIST datasets with cross-domain outliers. Our analysis demonstrates that both privacy mechanisms effectively map sensitive outliers to typical data points while maintaining utility in extreme Non-IID scenarios typical of cross-silo collaborations. Index Terms: Client-invariant Learning, Federated Learning (FL), Privacy-preserving Generative Models, Non-Independent and Identically Distributed (Non-IID), Heterogeneous Architectures

cs.LG↗

Rethinking Explanation Evaluation under the Retraining Scheme

Feature attribution has gained prominence as a tool for explaining model decisions, yet evaluating explanation quality remains challenging due to the absence of ground-truth explanations. To circumvent this, explanation-guided input manipulation has emerged as an indirect evaluation strategy, measuring explanation effectiveness through the impact of input modifications on model outcomes during inference. Despite the widespread use, a major concern with inference-based schemes is the distribution shift caused by such manipulations, which undermines the reliability of their assessments. The retraining-based scheme ROAR overcomes this issue by adapting the model to the altered data distribution. However, its evaluation results often contradict the theoretical foundations of widely accepted explainers. This work investigates this misalignment between empirical observations and theoretical expectations. In particular, we identify the sign issue as a key factor responsible for residual information that ultimately distorts retraining-based evaluation. Based on the analysis, we show that a straightforward reframing of the evaluation process can effectively resolve the identified issue. Building on the existing framework, we further propose novel variants that jointly structure a comprehensive perspective on explanation evaluation. These variants largely improve evaluation efficiency over the standard retraining protocol, thereby enhancing practical applicability for explainer selection and benchmarking. Following our proposed schemes, empirical results across various data scales provide deeper insights into the performance of carefully selected explainers, revealing open challenges and future directions in explainability research.

cs.LG↗

Model CBOR Serialization for Federated Learning

The typical federated learning workflow requires communication between a central server and a large set of clients synchronizing model parameters between each other. The current frameworks use communication protocols not suitable for resource-constrained devices and are either hard to deploy or require high-throughput links not available on these devices. In this paper, we present a generic message framework using CBOR for communication with existing federated learning frameworks optimised for use with resource-constrained devices and low power and lossy network links. We evaluate the resulting message sizes against JSON serialized messages where compare both with model parameters resulting in optimal and worst case serialization length, and with a real-world LeNet-5 model. Our benchmarks show that with our approach, messages are up to 75 % smaller in size when compared to the JSON alternative.

cs.NI↗