arXiv subjects

Subhankar Roy

Publications and source records attributed to Subhankar Roy.

At least 37 records · Page 2Linked to original sources

arXiv2025

A unique Neutrino Mass Matrix Texture under Exponential Parametrization

Under the exponential parametrization scheme $ m^ν_{ij}\sim r e^{i θ}$, we propose a \emph{unique} neutrino mass matrix texture that exhibits four correlations among its elements. The mixing scheme obtained from the proposed texture is consistent with experimental observations. We construct a concrete model based on the $SU(2)_L \times U(1)_Y \times A_4 \times Z_{10} \times Z_7$ group within the framework of the seesaw mechanism.

hep-ph↗

arXiv2025

Deviation from $μ-τ$ Symmetry with New Approaches

We propose two new predictive neutrino mass matrix textures that deviate from $μ$-$τ$ symmetry and explore their phenomenological implications. These textures are significant due to their predictive nature and the unique approach they introduce for breaking the $μ$-$τ$ symmetry. Both textures predict six physical parameters and impose sharp constraints on $θ_{23}$ and $δ$, considering both normal and inverted neutrino mass orderings. In addition, we realize the textures in their exact form within the framework of the seesaw mechanism, based on the $SU(2)_L \times A_4 \times Z_3$ symmetry group.

hep-ph↗

arXiv2025

Enhancing Plasticity for First Session Adaptation Continual Learning

The integration of large pre-trained models (PTMs) into Class-Incremental Learning (CIL) has facilitated the development of computationally efficient strategies such as First-Session Adaptation (FSA), which fine-tunes the model solely on the first task while keeping it frozen for subsequent tasks. Although effective in homogeneous task sequences, these approaches struggle when faced with the heterogeneity of real-world task distributions. We introduce Plasticity-Enhanced Test-Time Adaptation in Class-Incremental Learning (PLASTIC), a method that reinstates plasticity in CIL while preserving model stability. PLASTIC leverages Test-Time Adaptation (TTA) by dynamically fine-tuning LayerNorm parameters on unlabeled test data, enabling adaptability to evolving tasks and improving robustness against data corruption. To prevent TTA-induced model divergence and maintain stable learning across tasks, we introduce a teacher-student distillation framework, ensuring that adaptation remains controlled and generalizable. Extensive experiments across multiple benchmarks demonstrate that PLASTIC consistently outperforms both conventional and state-of-the-art PTM-based CIL approaches, while also exhibiting inherent robustness to data corruptions. Code is available at: https://github.com/IemProg/PLASTIC.

cs.CV↗

arXiv2025

Permuted Charged Lepton Correction in the Framework of Dirac Seesaw

A Dirac neutrino mass model is proposed, based on an extended group structure of $SU(2)_L \otimes U(1)_Y \otimes A_4 \otimes Z_3 \otimes Z_{10}$ with the Type-I seesaw mechanism. This work explores the impact of parametrization and permutation in the charged lepton diagonalizing matrix, driven by free parameters in the charged lepton sector, on the predictions of observable parameters. Some interesting consequences on the neutrino mass hierarchies, mixing angles and the Dirac CP phase are observed. The framework also finds application in the study of charged lepton flavour violation and dark matter.

hep-ph↗

arXiv2025

Unlearning Personal Data from a Single Image

Machine unlearning aims to erase data from a model as if the latter never saw them during training. While existing approaches unlearn information from complete or partial access to the training data, this access can be limited over time due to privacy regulations. Currently, no setting or benchmark exists to probe the effectiveness of unlearning methods in such scenarios. To fill this gap, we propose a novel task we call One-Shot Unlearning of Personal Identities (1-SHUI) that evaluates unlearning models when the training data is not available. We focus on unlearning identity data, which is specifically relevant due to current regulations requiring personal data deletion after training. To cope with data absence, we expect users to provide a portraiting picture to aid unlearning. We design requests on CelebA, CelebA-HQ, and MUFAC with different unlearning set sizes to evaluate applicable methods in 1-SHUI. Moreover, we propose MetaUnlearn, an effective method that meta-learns to forget identities from a single image. Our findings indicate that existing approaches struggle when data availability is limited, especially when there is a dissimilarity between the provided samples and the training data. Source code available at https://github.com/tdemin16/one-shui.

cs.CV↗

arXiv2025

Revisiting the Dirac Nature of Neutrinos in the Light of $Δ(27)$ and Cyclic Symmetries

Amid the uncertainty regarding the fundamental nature of neutrinos, we adhere to the Dirac description, and construct a model in the framework of $Δ(27)$ symmetry. The model successfully accounts for the hierarchical patterns of both charged lepton and neutrino masses. The neutrino mass matrix exhibits four texture zeroes, and the associated mixing scheme aligns with the experimental data, notably controlled by a single parameter.

hep-ph↗

arXiv2024

Weighted Ensemble Models Are Strong Continual Learners

In this work, we study the problem of continual learning (CL) where the goal is to learn a model on a sequence of tasks, such that the data from the previous tasks becomes unavailable while learning on the current task data. CL is essentially a balancing act between being able to learn on the new task (i.e., plasticity) and maintaining the performance on the previously learned concepts (i.e., stability). Intending to address the stability-plasticity trade-off, we propose to perform weight-ensembling of the model parameters of the previous and current tasks. This weighted-ensembled model, which we call Continual Model Averaging (or CoMA), attains high accuracy on the current task by leveraging plasticity, while not deviating too far from the previous weight configuration, ensuring stability. We also propose an improved variant of CoMA, named Continual Fisher-weighted Model Averaging (or CoFiMA), that selectively weighs each parameter in the weights ensemble by leveraging the Fisher information of the weights of the model. Both variants are conceptually simple, easy to implement, and effective in attaining state-of-the-art performance on several standard CL benchmarks. Code is available at: https://github.com/IemProg/CoFiMA.

cs.LG↗

arXiv2024

Neutrino Mixing from a Fresh Perspective

We propose a neutrino mass matrix texture bearing a suitable correlation $m_{22}=-2\,m_{13}$ and study its phenomenological implications. In light of both normal and inverted hierarchies, the texture imposes specific bounds on some observational parameters. As a potential application, the prediction of effective Majorana neutrino mass $m_{ββ}$ is visualized for both hierarchies. To understand the proposed texture from the first principle, we incorporate the type-I+II seesaw mechanism in association with $A_4 \times Z_{10} \times Z_2$ group.

hep-ph↗

arXiv2024

Large-scale Pre-trained Models are Surprisingly Strong in Incremental Novel Class Discovery

Discovering novel concepts in unlabelled datasets and in a continuous manner is an important desideratum of lifelong learners. In the literature such problems have been partially addressed under very restricted settings, where novel classes are learned by jointly accessing a related labelled set (e.g., NCD) or by leveraging only a supervisedly pre-trained model (e.g., class-iNCD). In this work we challenge the status quo in class-iNCD and propose a learning paradigm where class discovery occurs continuously and truly unsupervisedly, without needing any related labelled set. In detail, we propose to exploit the richer priors from strong self-supervised pre-trained models (PTM). To this end, we propose simple baselines, composed of a frozen PTM backbone and a learnable linear classifier, that are not only simple to implement but also resilient under longer learning scenarios. We conduct extensive empirical evaluation on a multitude of benchmarks and show the effectiveness of our proposed baselines when compared with sophisticated state-of-the-art methods. The code is open source.

cs.CV↗

arXiv2024

A Realistic Neutrino mixing scheme arising from $A_4$ symmetry

We propose a unique lepton mixing scheme and its association with an exact hierarchy-philic neutrino mass matrix texture in the light of Type-I+Type-II seesaw mechanism under the framework of $A_4 \times Z_{10}$ discrete flavour symmetry. The proposed model successfully predicts the normal ordering of neutrino masses and the two Majorana phases. Additionally, the analysis extends to the effective Majorana neutrino mass, in the context of neutrinoless double beta\,($0νββ$)-decay.

hep-ph↗

arXiv2024

Unveiling Neutrino Mysteries with $Δ(27)$ Symmetry

An elegant model is proposed by extending the Standard Model using the $Δ(27)\times Z_3 \times Z_{10}$ symmetry within the framework of the Type-I + Type-II seesaw mechanism. This model is particularly noteworthy for its ability to restrict the atmospheric mixing angle, $θ_{23}$, to specific values, and provides an explanation for the observed hierarchy of charged lepton masses. The neutrino mass matrix texture defined by three real parameters, predicts the three neutrino mass eigenvalues and the two Majorana phases. Furthermore, the model is tested against the experimental results of neutrino-less double beta ($0νββ$) decay and charged lepton flavour violation (cLFV) experiments.

hep-ph↗

arXiv2024

Less is more: Summarizing Patch Tokens for efficient Multi-Label Class-Incremental Learning

Prompt tuning has emerged as an effective rehearsal-free technique for class-incremental learning (CIL) that learns a tiny set of task-specific parameters (or prompts) to instruct a pre-trained transformer to learn on a sequence of tasks. Albeit effective, prompt tuning methods do not lend well in the multi-label class incremental learning (MLCIL) scenario (where an image contains multiple foreground classes) due to the ambiguity in selecting the correct prompt(s) corresponding to different foreground objects belonging to multiple tasks. To circumvent this issue we propose to eliminate the prompt selection mechanism by maintaining task-specific pathways, which allow us to learn representations that do not interact with the ones from the other tasks. Since independent pathways in truly incremental scenarios will result in an explosion of computation due to the quadratically complex multi-head self-attention (MSA) operation in prompt tuning, we propose to reduce the original patch token embeddings into summarized tokens. Prompt tuning is then applied to these fewer summarized tokens to compute the final representation. Our proposed method Multi-Label class incremental learning via summarising pAtch tokeN Embeddings (MULTI-LANE) enables learning disentangled task-specific representations in MLCIL while ensuring fast inference. We conduct experiments in common benchmarks and demonstrate that our MULTI-LANE achieves a new state-of-the-art in MLCIL. Additionally, we show that MULTI-LANE is also competitive in the CIL setting. Source code available at https://github.com/tdemin16/multi-lane

cs.CV↗

arXiv2024

The $μ$-$τ$ mixed symmetry and neutrino mass matrix

We propose an elegant neutrino mass matrix texture entitled $μ$-$τ$ mixed symmetry highlighting two simple correlations among its elements and a detailed analysis is carried out to see its phenomenological implications. The proposed texture is motivated in the framework of Seesaw mechanism in association with $A_4$ symmetry

hep-ph↗

arXiv2024

Collaborating Foundation Models for Domain Generalized Semantic Segmentation

Domain Generalized Semantic Segmentation (DGSS) deals with training a model on a labeled source domain with the aim of generalizing to unseen domains during inference. Existing DGSS methods typically effectuate robust features by means of Domain Randomization (DR). Such an approach is often limited as it can only account for style diversification and not content. In this work, we take an orthogonal approach to DGSS and propose to use an assembly of CoLlaborative FOUndation models for Domain Generalized Semantic Segmentation (CLOUDS). In detail, CLOUDS is a framework that integrates FMs of various kinds: (i) CLIP backbone for its robust feature representation, (ii) generative models to diversify the content, thereby covering various modes of the possible target distribution, and (iii) Segment Anything Model (SAM) for iteratively refining the predictions of the segmentation model. Extensive experiments show that our CLOUDS excels in adapting from synthetic to real DGSS benchmarks and under varying weather conditions, notably outperforming prior methods by 5.6% and 6.7% on averaged miou, respectively. The code is available at : https://github.com/yasserben/CLOUDS

cs.CV↗

arXiv2024

Democratizing Fine-grained Visual Recognition with Large Language Models

Identifying subordinate-level categories from images is a longstanding task in computer vision and is referred to as fine-grained visual recognition (FGVR). It has tremendous significance in real-world applications since an average layperson does not excel at differentiating species of birds or mushrooms due to subtle differences among the species. A major bottleneck in developing FGVR systems is caused by the need of high-quality paired expert annotations. To circumvent the need of expert knowledge we propose Fine-grained Semantic Category Reasoning (FineR) that internally leverages the world knowledge of large language models (LLMs) as a proxy in order to reason about fine-grained category names. In detail, to bridge the modality gap between images and LLM, we extract part-level visual attributes from images as text and feed that information to a LLM. Based on the visual attributes and its internal world knowledge the LLM reasons about the subordinate-level category names. Our training-free FineR outperforms several state-of-the-art FGVR and language and vision assistant models and shows promise in working in the wild and in new domains where gathering expert annotation is arduous.

cs.CV↗

arXiv2023

Constrained Neutrino Mass Matrix and Majorana Phases

We endeavor to constrain the neutrino mass matrix on the phenomenological ground and procure model-independent textures by emphasizing on the simple linear relationships among the mass matrix elements. These simple textures predict the two Majorana phases. In this regard, two types of parametrization of neutrino mass matrix: general and exponential are employed. We obtain fifty-three predictive neutrino mass matrix textures, out of which twenty-eight are associated with the general parametrization, and the rest belong to the exponential one. Apart from Type-A/P textures, the rest deal with the prediction of a few other oscillation parameters as well. We try to realize the proposed textures in the light of $A_4$, $Δ\,(27)$ and $T_7$ symmetry groups.

hep-ph↗

arXiv2023

Confined Neutrino Oscillation

A thought experiment is designed to speculate on the neutrino being trapped in an impenetrable potential well. Considering both relativistic and non-relativistic scenarios, we delve into several interesting facets connected to the flavour oscillation.

physics.gen-ph↗

arXiv2023

The Unreasonable Effectiveness of Large Language-Vision Models for Source-free Video Domain Adaptation

Source-Free Video Unsupervised Domain Adaptation (SFVUDA) task consists in adapting an action recognition model, trained on a labelled source dataset, to an unlabelled target dataset, without accessing the actual source data. The previous approaches have attempted to address SFVUDA by leveraging self-supervision (e.g., enforcing temporal consistency) derived from the target data itself. In this work, we take an orthogonal approach by exploiting "web-supervision" from Large Language-Vision Models (LLVMs), driven by the rationale that LLVMs contain a rich world prior surprisingly robust to domain-shift. We showcase the unreasonable effectiveness of integrating LLVMs for SFVUDA by devising an intuitive and parameter-efficient method, which we name Domain Adaptation with Large Language-Vision models (DALL-V), that distills the world prior and complementary source model information into a student network tailored for the target. Despite the simplicity, DALL-V achieves significant improvement over state-of-the-art SFVUDA methods.

cs.CV↗