arXiv subjects

Abhishek Mandal

Publications and source records attributed to Abhishek Mandal.

11 recordsLinked to original sources

arXiv2026

From Tokens to Semantics: Leveraging Complementary Signals for Hallucination Detection in Black-Box LLMs

When LLMs support public-facing or high-stakes workflows, missed fabrications can harm users and institutions, while false alarms consume limited human-review capacity. When no trusted context or reference document is available, we study two signals accessible through black-box model APIs: semantic entropy, which measures disagreement among sampled response meanings, and uncertainty derived from token log-probabilities. Their failure modes can be complementary: semantic entropy becomes uninformative when responses form one semantic cluster, while token uncertainty can miss consistently confident errors. We extend token-based uncertainty detection by aggregating token-level signals across sampled responses through our TopK method, evaluate the hybrid CoCoA method, which combines target-response uncertainty with semantic dissimilarity, and propose and study two supervised methods: Gated, which routes single-cluster cases to an aggregated-token-feature classifier, and Stacked, which learns jointly from semantic uncertainty and broader token features. We evaluate seven benchmarks, including five public benchmarks (four text datasets and multimodal handwritten-cheque extraction) and two constructed benchmarks (Financial Summaries and Long-Text QA), using four language models. In our evaluation across models and datasets, Stacked gave the best performance in nearly half of the cases, while TopK and CoCoA remain competitive without supervised training labels, although their thresholds require careful calibration. No method is universally strongest. We therefore evaluate performance at false-positive-rate budgets from 1% to 15%, assess their sensitivity to generation and calibration choices, and examine variation across dataset characteristics.

cs.CL↗

arXiv2026

Beyond Accuracy: Measuring Bias Acknowledgment in Chain-of-Thought Reasoning for Responsible AI Evaluation

Reasoning models are increasingly used in settings where the final answer is not the only object of review: educational tools may show students intermediate steps, decision-support systems may require human oversight, and audit workflows may inspect traces for misleading or biased input. In such settings, two responses can receive the same final-answer score while differing in whether the trace explicitly flags injected biasing content. Accuracy-only evaluation collapses these cases. We study this gap as a measurement blind spot for responsible evaluation and introduce a minimal trace-level diagnostic with two axes: \emph{susceptibility} (whether the bias breaks a previously correct answer) and \emph{acknowledgment} (whether the trace contains a rubric-defined surface reference to the injected content). Across thousands of biased GSM8K trials, GPT-4o and Claude Sonnet~4 have similar susceptibility rates ($1.3\%$ vs. $1.2\%$) but substantially different acknowledgment rates ($13.0\%$ vs. $75.0\%$) under the same rubric.

cs.LG↗

arXiv2026

Engineering of Tunable Topological Texture Transformation in Optical Skyrmions and Bimerons using Enantiomeric Excess

Optical skyrmions, which are the topologically protected quasiparticles and characterized by the nontrivial polarization textures, have emerged as a promising candidate due to their potential applications in optical communication, data storage, and particle manipulation. In this article, we propose and experimentally demonstrate an efficient and tunable approach for the dynamic transformation of generalized optical skyrmionic textures through the interaction of structured vector vortex beams with chiral media. By controlling the enantiomeric excess of an optically active material, we achieve on-demand conversion among Bloch, Neel or any intermediate skyrmionic states, extending also to optical bimerons. The topological conservation of the skyrmion number proves its robustness towards even higher-order textures. While maintaining a common path and stable setup, the proposed methodology provides an efficient and cost-effective approach towards the flexible manipulation of the topological textures, paving the way towards the understanding of topological transformation and engineering optical skyrmions for information processing or particle manipulation.

physics.optics↗

arXiv2026

Optimality of Sequential Filtering Under Independent Cost and Selectivity Models

Sequential filtering pipelines are a common design pattern in large-scale systems, where a large population of items is progressively reduced by a sequence of stages that each incur cost. Despite their prevalence in ranking systems, cascaded machine learning inference, and fraud detection, filter ordering is often determined by heuristics without formal guarantees. We formalize sequential filtering under an expected-cost objective and prove that, under an independence model, ordering filters by increasing ratio of cost to rejection probability minimizes expected total cost. Extensive Monte Carlo simulations show that the optimal ordering strictly dominates common heuristics across all runs, both in expectation and across the full distribution of outcomes.

cs.LG↗

arXiv2025

Tripartite Entanglement Generation in Atom-Coupled Dual Microresonators System

In this work, we investigate the emergence and control of genuine tripartite entanglement in a hybrid cavity quantum electrodynamics architecture consisting of two linearly coupled single mode resonators, one of which interacts coherently with a two level atom. An analytical framework is developed in a weak driving regime, where the system dynamically supports a delocalized hybrid excitation shared by the two photonic modes and the atomic degree of freedom. Tripartite concurrence fill has been used to characterize and identify parameter regimes of maximal multipartite quantum correlation that can be generated in this model. Additionally, we demonstrate how dissipative rates and detuning asymmetries govern the conversion of bipartite entanglement into a genuinely tripartite state, establishing a controllable transition from localized Jaynes Cummings correlations to delocalized photonic atomic entanglement networks. These findings outline a clear route to engineering steady state multipartite quantum resources in coupled cavity QED platforms, with direct relevance to quantum networking, distributed quantum information processing, and photonic state routing in scalable quantum architectures.

quant-ph↗

arXiv2024

Generated Bias: Auditing Internal Bias Dynamics of Text-To-Image Generative Models

Text-To-Image (TTI) Diffusion Models such as DALL-E and Stable Diffusion are capable of generating images from text prompts. However, they have been shown to perpetuate gender stereotypes. These models process data internally in multiple stages and employ several constituent models, often trained separately. In this paper, we propose two novel metrics to measure bias internally in these multistage multimodal models. Diffusion Bias was developed to detect and measures bias introduced by the diffusion stage of the models. Bias Amplification measures amplification of bias during the text-to-image conversion process. Our experiments reveal that TTI models amplify gender bias, the diffusion process itself contributes to bias and that Stable Diffusion v2 is more prone to gender bias than DALL-E 2.

cs.CV↗

arXiv2024

Flexible survival regression with variable selection for heterogeneous population

Survival regression is widely used to model time-to-events data, to explore how covariates may influence the occurrence of events. Modern datasets often encompass a vast number of covariates across many subjects, with only a subset of the covariates significantly affecting survival. Additionally, subjects often belong to an unknown number of latent groups, where covariate effects on survival differ significantly across groups. The proposed methodology addresses both challenges by simultaneously identifying the latent sub-groups in the heterogeneous population and evaluating covariate significance within each sub-group. This approach is shown to enhance the predictive accuracy for time-to-event outcomes, via uncovering varying risk profiles within the underlying heterogeneous population and is thereby helpful to device targeted disease management strategies.

stat.ME↗

arXiv2023

Biased Attention: Do Vision Transformers Amplify Gender Bias More than Convolutional Neural Networks?

Deep neural networks used in computer vision have been shown to exhibit many social biases such as gender bias. Vision Transformers (ViTs) have become increasingly popular in computer vision applications, outperforming Convolutional Neural Networks (CNNs) in many tasks such as image classification. However, given that research on mitigating bias in computer vision has primarily focused on CNNs, it is important to evaluate the effect of a different network architecture on the potential for bias amplification. In this paper we therefore introduce a novel metric to measure bias in architectures, Accuracy Difference. We examine bias amplification when models belonging to these two architectures are used as a part of large multimodal models, evaluating the different image encoders of Contrastive Language Image Pretraining which is an important model used in many generative models such as DALL-E and Stable Diffusion. Our experiments demonstrate that architecture can play a role in amplifying social biases due to the different techniques employed by the models for feature extraction and embedding as well as their different learning properties. This research found that ViTs amplified gender bias to a greater extent than CNNs

cs.CV↗

arXiv2023

Gender Bias in Multimodal Models: A Transnational Feminist Approach Considering Geographical Region and Culture

Deep learning based visual-linguistic multimodal models such as Contrastive Language Image Pre-training (CLIP) have become increasingly popular recently and are used within text-to-image generative models such as DALL-E and Stable Diffusion. However, gender and other social biases have been uncovered in these models, and this has the potential to be amplified and perpetuated through AI systems. In this paper, we present a methodology for auditing multimodal models that consider gender, informed by concepts from transnational feminism, including regional and cultural dimensions. Focusing on CLIP, we found evidence of significant gender bias with varying patterns across global regions. Harmful stereotypical associations were also uncovered related to visual cultural cues and labels such as terrorism. Levels of gender bias uncovered within CLIP for different regions aligned with global indices of societal gender equality, with those from the Global South reflecting the highest levels of gender bias.

cs.CY↗

arXiv2023

Understanding EEG signals for subject-wise Definition of Armoni Activities

In a growing world of technology, psychological disorders became a challenge to be solved. The methods used for cognitive stimulation are very conventional and based on one-way communication, which only relies on the material or method used for training of an individual. It doesn't use any kind of feedback from the individual to analyze the progress of the training process. We have proposed a closed-loop methodology to improve the cognitive state of a person with ID (Intellectual disability). We have used a platform named 'Armoni', for providing training to the intellectually disabled individuals. The learning is performed in a closed-loop by using feedback in the form of change in affective state. For feedback to the Armoni, an EEG (Electroencephalograph) headband is used. All the changes in EEG are observed and classified against the change in the mean and standard deviation value of all frequency bands of signal. This comparison is being helpful in defining every activity with respect to change in brain signals. In this paper, we have discussed the process of treatment of EEG signal and its definition against the different activities of Armoni. We have tested it on 6 different systems with different age groups and cognitive levels.

eess.SP↗

arXiv2023

Multimodal Composite Association Score: Measuring Gender Bias in Generative Multimodal Models

Generative multimodal models based on diffusion models have seen tremendous growth and advances in recent years. Models such as DALL-E and Stable Diffusion have become increasingly popular and successful at creating images from texts, often combining abstract ideas. However, like other deep learning models, they also reflect social biases they inherit from their training data, which is often crawled from the internet. Manually auditing models for biases can be very time and resource consuming and is further complicated by the unbounded and unconstrained nature of inputs these models can take. Research into bias measurement and quantification has generally focused on small single-stage models working on a single modality. Thus the emergence of multistage multimodal models requires a different approach. In this paper, we propose Multimodal Composite Association Score (MCAS) as a new method of measuring gender bias in multimodal generative models. Evaluating both DALL-E 2 and Stable Diffusion using this approach uncovered the presence of gendered associations of concepts embedded within the models. We propose MCAS as an accessible and scalable method of quantifying potential bias for models with different modalities and a range of potential biases.

cs.CV↗