arXiv subjects

Abhik Bhattacharjee

Publications and source records attributed to Abhik Bhattacharjee.

13 recordsLinked to original sources

arXiv2026

Transient Early Dark Energy-Like Dynamics as a Mechanism for Enhanced Early Structure Formation in the JWST Era

The discovery of massive galaxies at redshifts $z\gtrsim10$ by the James Webb Space Telescope (JWST) has renewed interest in cosmological mechanisms capable of enhancing early structure formation while preserving the successful large-scale predictions of the standard $\Lambda$CDM model. We investigate a phenomenological scenario in which an exotic dark matter species can undergo a transient early dark energy-like phase during the radiation-dominated era ($10^{-7}\lesssim a\lesssim10^{-5}$) before reverting to pressureless cold dark matter. We utilize the generalized dark matter framework to model this species, which is restricted to a sub-percent fraction of the total dark matter component by CMB, BAO and Type Ia supernova data. Its background and perturbation dynamics are characterized by a time-dependent equation of state, $w(a)$, and a time- as well as scale-dependent sound speed, $c_s^2(a,k)$. The temporary negative equation of state, combined with our phenomenological pressure-response prescription, induces a finite interval of negative effective sound speed squared. This triggers an instability-driven growth of density perturbations over a limited range of comoving scales, thereby enhancing the formation of early dark matter halos. We find that the enhanced halo abundance can substantially reduce the star-formation efficiencies required to reproduce the observed abundance of JWST galaxies relative to the standard $\Lambda$CDM scenario, especially at higher redshifts. Our results demonstrate that transient early dark energy-like dynamics in a subdominant dark matter component provide a viable mechanism for enhancing early structure formation and offer a new framework for interpreting the abundance of high-redshift galaxies observed by JWST and future surveys.

astro-ph.CO↗

arXiv2025

Forecasting Constraints on Non-Thermal Light Massive Relics from Future CMB Experiments (CMB-S4/Simons Observatory)

In this work we present Fisher forecasts on \textit{non-thermal LiMR} models for a CMB Stage IV-like experiment and the Simons Observatory -- particularly focusing on a model of inflaton/moduli decay giving rise to non-thermally distributed dark sector particles, and also comparing our results with those for sterile particles following the Dodelson-Widrow distribution. Two independent parameters, $\Delta N_\mathrm{eff}$ and $M_\mathrm{sp}^\mathrm{eff}$, influence linear cosmological observables. We find $\Delta N_\mathrm{eff}$ to be more tightly constrained (by a factor of $10$) for a less abundant, heavier LiMR which becomes fully non-relativistic around matter-radiation equality than a more abundant, lighter LiMR which becomes fully non-relativistic just after recombination. The uncertainties on $M_\mathrm{sp}^\mathrm{eff}$ differ by a factor of $\sim3$ between the two cases. Our analysis also reveals distinct parameter correlations: the phenomenological parameters $\{\Delta N_\mathrm{eff},M_\mathrm{sp}^\mathrm{eff}\}$ are found to be negatively correlated for the former case and positively correlated for the latter. We obtain similar projected uncertainties on the cosmological parameters (in either case) for both the inflaton/moduli decay and the Dodelson-Widrow models when the first two moments of the LiMR distribution function, related to the phenomenological parameters, are matched. Finally, by constructing a modified distribution that matches the first two moments of the Dodelson-Widrow but deviates maximally in the third moment, we demonstrate that CMB Stage IV data is not expected to be sensitive to higher moments of the distribution.

astro-ph.CO↗

arXiv2024

Multi-ToM: Evaluating Multilingual Theory of Mind Capabilities in Large Language Models

Theory of Mind (ToM) refers to the cognitive ability to infer and attribute mental states to oneself and others. As large language models (LLMs) are increasingly evaluated for social and cognitive capabilities, it remains unclear to what extent these models demonstrate ToM across diverse languages and cultural contexts. In this paper, we introduce a comprehensive study of multilingual ToM capabilities aimed at addressing this gap. Our approach includes two key components: (1) We translate existing ToM datasets into multiple languages, effectively creating a multilingual ToM dataset and (2) We enrich these translations with culturally specific elements to reflect the social and cognitive scenarios relevant to diverse populations. We conduct extensive evaluations of six state-of-the-art LLMs to measure their ToM performance across both the translated and culturally adapted datasets. The results highlight the influence of linguistic and cultural diversity on the models' ability to exhibit ToM, and questions their social reasoning capabilities. This work lays the groundwork for future research into enhancing LLMs' cross-cultural social cognition and contributes to the development of more culturally aware and socially intelligent AI systems. All our data and code are publicly available.

cs.CL↗

arXiv2024

Revisiting holographic codes with fractal-like boundary erasures

In this paper we investigate the code properties of holographic fractal geometries initiated in \cite{Pastawski:2016qrs}. We study reconstruction wedges in $AdS_3/CFT_2$ for black hole backgrounds, which are in qualitative agreement with the vacuum-AdS approximation using generalized entanglement entropy in \cite{Bao:2022tgv}. In higher dimensions, we study reconstruction wedges for the infinite, straight strip in $AdS_{d+1}/CFT_{d}$ and clarify the roles of `straight' and `infinite' in their code properties. Lastly, we comment on uberholography from the perspective of complexity transfer and one-shot holography.

hep-th↗

arXiv2024

IllusionVQA: A Challenging Optical Illusion Dataset for Vision Language Models

The advent of Vision Language Models (VLM) has allowed researchers to investigate the visual understanding of a neural network using natural language. Beyond object classification and detection, VLMs are capable of visual comprehension and common-sense reasoning. This naturally led to the question: How do VLMs respond when the image itself is inherently unreasonable? To this end, we present IllusionVQA: a diverse dataset of challenging optical illusions and hard-to-interpret scenes to test the capability of VLMs in two distinct multiple-choice VQA tasks - comprehension and soft localization. GPT4V, the best performing VLM, achieves 62.99% accuracy (4-shot) on the comprehension task and 49.7% on the localization task (4-shot and Chain-of-Thought). Human evaluation reveals that humans achieve 91.03% and 100% accuracy in comprehension and localization. We discover that In-Context Learning (ICL) and Chain-of-Thought reasoning substantially degrade the performance of Gemini-Pro in the localization task. Tangentially, we discover a potential weakness in the ICL capabilities of VLMs: they fail to locate optical illusions even when the correct answer is in the context window as a few-shot example.

cs.CV↗

arXiv2024

An Empirical Study on the Characteristics of Bias upon Context Length Variation for Bangla

Pretrained language models inherently exhibit various social biases, prompting a crucial examination of their social impact across various linguistic contexts due to their widespread usage. Previous studies have provided numerous methods for intrinsic bias measurements, predominantly focused on high-resource languages. In this work, we aim to extend these investigations to Bangla, a low-resource language. Specifically, in this study, we (1) create a dataset for intrinsic gender bias measurement in Bangla, (2) discuss necessary adaptations to apply existing bias measurement methods for Bangla, and (3) examine the impact of context length variation on bias measurement, a factor that has been overlooked in previous studies. Through our experiments, we demonstrate a clear dependency of bias metrics on context length, highlighting the need for nuanced considerations in Bangla bias analysis. We consider our work as a stepping stone for bias measurement in the Bangla Language and make all of our resources publicly available to support future research.

cs.CL↗

arXiv2023

CrossSum: Beyond English-Centric Cross-Lingual Summarization for 1,500+ Language Pairs

We present CrossSum, a large-scale cross-lingual summarization dataset comprising 1.68 million article-summary samples in 1,500+ language pairs. We create CrossSum by aligning parallel articles written in different languages via cross-lingual retrieval from a multilingual abstractive summarization dataset and perform a controlled human evaluation to validate its quality. We propose a multistage data sampling algorithm to effectively train a cross-lingual summarization model capable of summarizing an article in any target language. We also introduce LaSE, an embedding-based metric for automatically evaluating model-generated summaries. LaSE is strongly correlated with ROUGE and, unlike ROUGE, can be reliably measured even in the absence of references in the target language. Performance on ROUGE and LaSE indicate that our proposed model consistently outperforms baseline models. To the best of our knowledge, CrossSum is the largest cross-lingual summarization dataset and the first ever that is not centered around English. We are releasing the dataset, training and evaluation scripts, and models to spur future research on cross-lingual summarization. The resources can be found at https://github.com/csebuetnlp/CrossSum

cs.CL↗

arXiv2023

BanglaNLG and BanglaT5: Benchmarks and Resources for Evaluating Low-Resource Natural Language Generation in Bangla

This work presents BanglaNLG, a comprehensive benchmark for evaluating natural language generation (NLG) models in Bangla, a widely spoken yet low-resource language. We aggregate six challenging conditional text generation tasks under the BanglaNLG benchmark, introducing a new dataset on dialogue generation in the process. Furthermore, using a clean corpus of 27.5 GB of Bangla data, we pretrain BanglaT5, a sequence-to-sequence Transformer language model for Bangla. BanglaT5 achieves state-of-the-art performance in all of these tasks, outperforming several multilingual models by up to 9% absolute gain and 32% relative gain. We are making the new dialogue dataset and the BanglaT5 model publicly available at https://github.com/csebuetnlp/BanglaNLG in the hope of advancing future research on Bangla NLG.

cs.CL↗

arXiv2022

BanglaParaphrase: A High-Quality Bangla Paraphrase Dataset

In this work, we present BanglaParaphrase, a high-quality synthetic Bangla Paraphrase dataset curated by a novel filtering pipeline. We aim to take a step towards alleviating the low resource status of the Bangla language in the NLP domain through the introduction of BanglaParaphrase, which ensures quality by preserving both semantics and diversity, making it particularly useful to enhance other Bangla datasets. We show a detailed comparative analysis between our dataset and models trained on it with other existing works to establish the viability of our synthetic paraphrase data generation pipeline. We are making the dataset and models publicly available at https://github.com/csebuetnlp/banglaparaphrase to further the state of Bangla NLP.

cs.CL↗

arXiv2022

GEMv2: Multilingual NLG Benchmarking in a Single Line of Code

Evaluation in machine learning is usually informed by past choices, for example which datasets or metrics to use. This standardization enables the comparison on equal footing using leaderboards, but the evaluation choices become sub-optimal as better alternatives arise. This problem is especially pertinent in natural language generation which requires ever-improving suites of datasets, metrics, and human evaluation to make definitive claims. To make following best model evaluation practices easier, we introduce GEMv2. The new version of the Generation, Evaluation, and Metrics Benchmark introduces a modular infrastructure for dataset, model, and metric developers to benefit from each others work. GEMv2 supports 40 documented datasets in 51 languages. Models for all datasets can be evaluated online and our interactive data card creation and rendering tools make it easier to add new datasets to the living benchmark.

cs.CL↗

arXiv2022

BanglaBERT: Language Model Pretraining and Benchmarks for Low-Resource Language Understanding Evaluation in Bangla

In this work, we introduce BanglaBERT, a BERT-based Natural Language Understanding (NLU) model pretrained in Bangla, a widely spoken yet low-resource language in the NLP literature. To pretrain BanglaBERT, we collect 27.5 GB of Bangla pretraining data (dubbed `Bangla2B+') by crawling 110 popular Bangla sites. We introduce two downstream task datasets on natural language inference and question answering and benchmark on four diverse NLU tasks covering text classification, sequence labeling, and span prediction. In the process, we bring them under the first-ever Bangla Language Understanding Benchmark (BLUB). BanglaBERT achieves state-of-the-art results outperforming multilingual and monolingual models. We are making the models, datasets, and a leaderboard publicly available at https://github.com/csebuetnlp/banglabert to advance Bangla NLP.

cs.CL↗

arXiv2021

XL-Sum: Large-Scale Multilingual Abstractive Summarization for 44 Languages

Contemporary works on abstractive text summarization have focused primarily on high-resource languages like English, mostly due to the limited availability of datasets for low/mid-resource ones. In this work, we present XL-Sum, a comprehensive and diverse dataset comprising 1 million professionally annotated article-summary pairs from BBC, extracted using a set of carefully designed heuristics. The dataset covers 44 languages ranging from low to high-resource, for many of which no public dataset is currently available. XL-Sum is highly abstractive, concise, and of high quality, as indicated by human and intrinsic evaluation. We fine-tune mT5, a state-of-the-art pretrained multilingual model, with XL-Sum and experiment on multilingual and low-resource summarization tasks. XL-Sum induces competitive results compared to the ones obtained using similar monolingual datasets: we show higher than 11 ROUGE-2 scores on 10 languages we benchmark on, with some of them exceeding 15, as obtained by multilingual training. Additionally, training on low-resource languages individually also provides competitive performance. To the best of our knowledge, XL-Sum is the largest abstractive summarization dataset in terms of the number of samples collected from a single source and the number of languages covered. We are releasing our dataset and models to encourage future research on multilingual abstractive summarization. The resources can be found at \url{https://github.com/csebuetnlp/xl-sum}.

cs.CL↗

arXiv2020

Not Low-Resource Anymore: Aligner Ensembling, Batch Filtering, and New Datasets for Bengali-English Machine Translation

Despite being the seventh most widely spoken language in the world, Bengali has received much less attention in machine translation literature due to being low in resources. Most publicly available parallel corpora for Bengali are not large enough; and have rather poor quality, mostly because of incorrect sentence alignments resulting from erroneous sentence segmentation, and also because of a high volume of noise present in them. In this work, we build a customized sentence segmenter for Bengali and propose two novel methods for parallel corpus creation on low-resource setups: aligner ensembling and batch filtering. With the segmenter and the two methods combined, we compile a high-quality Bengali-English parallel corpus comprising of 2.75 million sentence pairs, more than 2 million of which were not available before. Training on neural models, we achieve an improvement of more than 9 BLEU score over previous approaches to Bengali-English machine translation. We also evaluate on a new test set of 1000 pairs made with extensive quality control. We release the segmenter, parallel corpus, and the evaluation set, thus elevating Bengali from its low-resource status. To the best of our knowledge, this is the first ever large scale study on Bengali-English machine translation. We believe our study will pave the way for future research on Bengali-English machine translation as well as other low-resource languages. Our data and code are available at https://github.com/csebuetnlp/banglanmt.

cs.CL↗