Sayan Nag

Mona Gandhi, KJ Joseph, Srinivasan Parthasarathy et al.

Effective visual communication stems from the harmony of multiple design principles, such as readability, contrast, alignment, overlap, and coherence, which collectively govern clarity and intent of the communicator. While human designers reason holistically over these principles, machine agents typically condense them into a single heuristic score, offering limited interpretability and diagnostic precision. To address this gap, we introduce PRISM (PRinciple-aware, Interpretable, and Structure-guided Design Modifications), a benchmark that systematically perturbs professional layouts from the Crello dataset along measurable design principles. The benchmark comprises 100K perturbed training samples and 10K perturbed validation designs, each isolating a specific principle violation for controlled analysis of multimodal reasoning about design quality. We show that models like Qwen-2.5-VL and GPT-4o-mini are largely insensitive to targeted principle degradations, whereas GPT-4o exhibits global awareness without fine-grained disentanglement. Building on these insights, we propose a multi-scale evaluation framework that integrates lightweight scorers for quantitative assessment, instruction-tuned vision-language models for localised feedback, and prompt-based methods for global reasoning. Our framework provides interpretable explanations of design failures. Using these localised insights, we show targeted refinements that improve layout quality. Together, PRISM and our framework lay the foundation for interpretable design-literate multimodal reasoning systems.

Shraman Pramanick, Li Jing, Sayan Nag et al. · openai

Vision-language pre-training (VLP) has recently proven highly effective for various uni- and multi-modal downstream applications. However, most existing end-to-end VLP methods use high-resolution image-text box data to perform well on fine-grained region-level tasks, such as object detection, segmentation, and referring expression comprehension. Unfortunately, such high-resolution images with accurate bounding box annotations are expensive to collect and use for supervision at scale. In this work, we propose VoLTA (Vision-Language Transformer with weakly-supervised local-feature Alignment), a new VLP paradigm that only utilizes image-caption data but achieves fine-grained region-level image understanding, eliminating the use of expensive box annotations. VoLTA adopts graph optimal transport-based weakly-supervised alignment on local image patches and text tokens to germinate an explicit, self-normalized, and interpretable low-level matching criterion. In addition, VoLTA pushes multi-modal fusion deep into the uni-modal backbones during pre-training and removes fusion-specific transformer layers, further reducing memory requirements. Extensive experiments on a wide range of vision- and vision-language downstream tasks demonstrate the effectiveness of VoLTA on fine-grained applications without compromising the coarse-grained downstream performance, often outperforming methods using significantly more caption and box annotations.

Shraman Pramanick, Yale Song, Sayan Nag et al. · microsoft-research, uw

Video-language pre-training (VLP) has become increasingly important due to its ability to generalize to various vision and language tasks. However, existing egocentric VLP frameworks utilize separate video and language encoders and learn task-specific cross-modal information only during fine-tuning, limiting the development of a unified system. In this work, we introduce the second generation of egocentric video-language pre-training (EgoVLPv2), a significant improvement from the previous generation, by incorporating cross-modal fusion directly into the video and language backbones. EgoVLPv2 learns strong video-text representation during pre-training and reuses the cross-modal attention modules to support different downstream tasks in a flexible and efficient manner, reducing fine-tuning costs. Moreover, our proposed fusion in the backbone strategy is more lightweight and compute-efficient than stacking additional fusion-specific layers. Extensive experiments on a wide range of VL tasks demonstrate the effectiveness of EgoVLPv2 by achieving consistent state-of-the-art performance over strong baselines across all downstream. Our project page can be found at https://shramanpramanick.github.io/EgoVLPv2/.

Hritam Basak, Soumitri Chattopadhyay, Rohit Kundu et al.

Due to the scarcity of labeled data, Contrastive Self-Supervised Learning (SSL) frameworks have lately shown great potential in several medical image analysis tasks. However, the existing contrastive mechanisms are sub-optimal for dense pixel-level segmentation tasks due to their inability to mine local features. To this end, we extend the concept of metric learning to the segmentation task, using a dense (dis)similarity learning for pre-training a deep encoder network, and employing a semi-supervised paradigm to fine-tune for the downstream task. Specifically, we propose a simple convolutional projection head for obtaining dense pixel-level features, and a new contrastive loss to utilize these dense projections thereby improving the local representations. A bidirectional consistency regularization mechanism involving two-stream model training is devised for the downstream task. Upon comparison, our IDEAL method outperforms the SoTA methods by fair margins on cardiac MRI segmentation. Code available: https://github.com/hritam-98/IDEAL-ICASSP23

Sanjoy Chowdhury, Sayan Nag, Subhrajyoti Dasgupta et al.

Leveraging Large Language Models' remarkable proficiency in text-based tasks, recent works on Multi-modal LLMs (MLLMs) extend them to other modalities like vision and audio. However, the progress in these directions has been mostly focused on tasks that only require a coarse-grained understanding of the audio-visual semantics. We present Meerkat, an audio-visual LLM equipped with a fine-grained understanding of image and audio both spatially and temporally. With a new modality alignment module based on optimal transport and a cross-attention module that enforces audio-visual consistency, Meerkat can tackle challenging tasks such as audio referred image grounding, image guided audio temporal localization, and audio-visual fact-checking. Moreover, we carefully curate a large dataset AVFIT that comprises 3M instruction tuning samples collected from open-source datasets, and introduce MeerkatBench that unifies five challenging audio-visual tasks. We achieve state-of-the-art performance on all these downstream tasks with a relative improvement of up to 37.12%.

Sayan Nag, Koustava Goswami, Srikrishna Karanam

Referring Expression Segmentation (RES) aims to provide a segmentation mask of the target object in an image referred to by the text (i.e., referring expression). Existing methods require large-scale mask annotations. Moreover, such approaches do not generalize well to unseen/zero-shot scenarios. To address the aforementioned issues, we propose a weakly-supervised bootstrapping architecture for RES with several new algorithmic innovations. To the best of our knowledge, ours is the first approach that considers only a fraction of both mask and box annotations (shown in Figure 1 and Table 1) for training. To enable principled training of models in such low-annotation settings, improve image-text region-level alignment, and further enhance spatial localization of the target object in the image, we propose Cross-modal Fusion with Attention Consistency module. For automatic pseudo-labeling of unlabeled samples, we introduce a novel Mask Validity Filtering routine based on a spatially aware zero-shot proposal scoring approach. Extensive experiments show that with just 30% annotations, our model SafaRi achieves 59.31 and 48.26 mIoUs as compared to 58.93 and 48.19 mIoUs obtained by the fully-supervised SOTA method SeqTR respectively on RefCOCO+@testA and RefCOCO+testB datasets. SafaRi also outperforms SeqTR by 11.7% (on RefCOCO+testA) and 19.6% (on RefCOCO+testB) in a fully-supervised setting and demonstrates strong generalization capabilities in unseen/zero-shot tasks.

Arman Zarei, Samyadeep Basu, Mobina Pournemat et al.

Instruction-based image editing models have recently achieved impressive performance, enabling complex edits to an input image from a multi-instruction prompt. However, these models apply each instruction in the prompt with a fixed strength, limiting the user's ability to precisely and continuously control the intensity of individual edits. We introduce SliderEdit, a framework for continuous image editing with fine-grained, interpretable instruction control. Given a multi-part edit instruction, SliderEdit disentangles the individual instructions and exposes each as a globally trained slider, allowing smooth adjustment of its strength. Unlike prior works that introduced slider-based attribute controls in text-to-image generation, typically requiring separate training or fine-tuning for each attribute or concept, our method learns a single set of low-rank adaptation matrices that generalize across diverse edits, attributes, and compositional instructions. This enables continuous interpolation along individual edit dimensions while preserving both spatial locality and global semantic consistency. We apply SliderEdit to state-of-the-art image editing models, including FLUX-Kontext and Qwen-Image-Edit, and observe substantial improvements in edit controllability, visual consistency, and user steerability. To the best of our knowledge, we are the first to explore and propose a framework for continuous, fine-grained instruction control in instruction-based image editing models. Our results pave the way for interactive, instruction-driven image manipulation with continuous and compositional control.

Soumitri Chattopadhyay, Soham Ganguly, Sreejit Chaudhury et al.

The success of self-supervised learning (SSL) has mostly been attributed to the availability of unlabeled yet large-scale datasets. However, in a specialized domain such as medical imaging which is a lot different from natural images, the assumption of data availability is unrealistic and impractical, as the data itself is scanty and found in small databases, collected for specific prognosis tasks. To this end, we seek to investigate the applicability of self-supervised learning algorithms on small-scale medical imaging datasets. In particular, we evaluate $4$ state-of-the-art SSL methods on three publicly accessible \emph{small} medical imaging datasets. Our investigation reveals that in-domain low-resource SSL pre-training can yield competitive performance to transfer learning from large-scale datasets (such as ImageNet). Furthermore, we extensively analyse our empirical findings to provide valuable insights that can motivate for further research towards circumventing the need for pre-training on a large image corpus. To the best of our knowledge, this is the first attempt to holistically explore self-supervision on low-resource medical datasets.

Ahana Deb, Sayan Nag, Ayan Mahapatra et al.

Spoken languages often utilise intonation, rhythm, intensity, and structure, to communicate intention, which can be interpreted differently depending on the rhythm of speech of their utterance. These speech acts provide the foundation of communication and are unique in expression to the language. Recent advancements in attention-based models, demonstrating their ability to learn powerful representations from multilingual datasets, have performed well in speech tasks and are ideal to model specific tasks in low resource languages. Here, we develop a novel multimodal approach combining two models, wav2vec2.0 for audio and MarianMT for text translation, by using multimodal attention fusion to predict speech acts in our prepared Bengali speech corpus. We also show that our model BeAts ($\underline{\textbf{Be}}$ngali speech acts recognition using Multimodal $\underline{\textbf{At}}$tention Fu$\underline{\textbf{s}}$ion) significantly outperforms both the unimodal baseline using only speech data and a simpler bimodal fusion using both speech and text data. Project page: https://soumitri2001.github.io/BeAts

Soumitri Chattopadhyay, Soham Ganguly, Sreejit Chaudhury et al.

Privacy and annotation bottlenecks are two major issues that profoundly affect the practicality of machine learning-based medical image analysis. Although significant progress has been made in these areas, these issues are not yet fully resolved. In this paper, we seek to tackle these concerns head-on and systematically explore the applicability of non-contrastive self-supervised learning (SSL) algorithms under federated learning (FL) simulations for medical image analysis. We conduct thorough experimentation of recently proposed state-of-the-art non-contrastive frameworks under standard FL setups. With the SoTA Contrastive Learning algorithm, SimCLR as our comparative baseline, we benchmark the performances of our 4 chosen non-contrastive algorithms under non-i.i.d. data conditions and with a varying number of clients. We present a holistic evaluation of these techniques on 6 standardized medical imaging datasets. We further analyse different trends inferred from the findings of our research, with the aim to find directions for further research based on ours. To the best of our knowledge, ours is the first to perform such a thorough analysis of federated self-supervised learning for medical imaging. All of our source code will be made public upon acceptance of the paper.

Sanjoy Chowdhury, Sayan Nag, Subhrajyoti Dasgupta et al.

With the rapid advancement of Multi-modal Large Language Models (MLLMs), several diagnostic benchmarks have recently been developed to assess these models' multi-modal reasoning proficiency. However, these benchmarks are restricted to assessing primarily the visual aspect and do not examine the holistic audio-visual (AV) understanding. Moreover, currently, there are no benchmarks that investigate the capabilities of AVLLMs to calibrate their responses when presented with perturbed inputs. To this end, we introduce Audio-Visual Trustworthiness assessment Benchmark (AVTrustBench), comprising 600K samples spanning over 9 meticulously crafted tasks, evaluating the capabilities of AVLLMs across three distinct dimensions: Adversarial attack, Compositional reasoning, and Modality-specific dependency. Using our benchmark we extensively evaluate 13 state-of-the-art AVLLMs. The findings reveal that the majority of existing models fall significantly short of achieving human-like comprehension, offering valuable insights for future research directions. To alleviate the limitations in the existing approaches, we further propose a robust, model-agnostic calibrated audio-visual preference optimization based training strategy CAVPref, obtaining a gain up to 30.19% across all 9 tasks. We will publicly release our code and benchmark to facilitate future research in this direction.

Sanjoy Chowdhury, Hanan Gani, Nishit Anand et al.

Recent advancements in reasoning optimization have greatly enhanced the performance of large language models (LLMs). However, existing work fails to address the complexities of audio-visual scenarios, underscoring the need for further research. In this paper, we introduce AURELIA, a novel actor-critic based audio-visual (AV) reasoning framework that distills structured, step-by-step reasoning into AVLLMs at test time, improving their ability to process complex multi-modal inputs without additional training or fine-tuning. To further advance AVLLM reasoning skills, we present AVReasonBench, a challenging benchmark comprising 4500 audio-visual questions, each paired with detailed step-by-step reasoning. Our benchmark spans six distinct tasks, including AV-GeoIQ, which evaluates AV reasoning combined with geographical and cultural knowledge. Evaluating 18 AVLLMs on AVReasonBench reveals significant limitations in their multi-modal reasoning capabilities. Using AURELIA, we achieve up to a 100% relative improvement, demonstrating its effectiveness. This performance gain highlights the potential of reasoning-enhanced data generation for advancing AVLLMs in real-world applications. Our code and data will be publicly released at: https: //github.com/schowdhury671/aurelia.

Siyuan Huang, Yutong Gao, Juyang Bai et al.

Creating high-quality figures and visualizations for scientific papers is a time-consuming task that requires both deep domain knowledge and professional design skills. Despite over 2.5 million scientific papers published annually, the figure generation process remains largely manual. We introduce $\textbf{SciFig}$, an end-to-end AI agent system that generates publication-ready pipeline figures directly from research paper texts. SciFig uses a hierarchical layout generation strategy, which parses research descriptions to identify component relationships, groups related elements into functional modules, and generates inter-module connections to establish visual organization. Furthermore, an iterative chain-of-thought (CoT) feedback mechanism progressively improves layouts through multiple rounds of visual analysis and reasoning. We introduce a rubric-based evaluation framework that analyzes 2,219 real scientific figures to extract evaluation rubrics and automatically generates comprehensive evaluation criteria. SciFig demonstrates remarkable performance: achieving 70.1$\%$ overall quality on dataset-level evaluation and 66.2$\%$ on paper-specific evaluation, and consistently high scores across metrics such as visual clarity, structural organization, and scientific accuracy. SciFig figure generation pipeline and our evaluation benchmark will be open-sourced.

Shraman Pramanick, Guangxing Han, Rui Hou et al.

The ability of large language models (LLMs) to process visual inputs has given rise to general-purpose vision systems, unifying various vision-language (VL) tasks by instruction tuning. However, due to the enormous diversity in input-output formats in the vision domain, existing general-purpose models fail to successfully integrate segmentation and multi-image inputs with coarse-level tasks into a single framework. In this work, we introduce VistaLLM, a powerful visual system that addresses coarse- and fine-grained VL tasks over single and multiple input images using a unified framework. VistaLLM utilizes an instruction-guided image tokenizer that filters global embeddings using task descriptions to extract compressed and refined features from numerous images. Moreover, VistaLLM employs a gradient-aware adaptive sampling technique to represent binary segmentation masks as sequences, significantly improving over previously used uniform sampling. To bolster the desired capability of VistaLLM, we curate CoinIt, a comprehensive coarse-to-fine instruction tuning dataset with 6.8M samples. We also address the lack of multi-image grounding datasets by introducing a novel task, AttCoSeg (Attribute-level Co-Segmentation), which boosts the model's reasoning and grounding capability over multiple input images. Extensive experiments on a wide range of V- and VL tasks demonstrate the effectiveness of VistaLLM by achieving consistent state-of-the-art performance over strong baselines across all downstream tasks. Our project page can be found at https://shramanpramanick.github.io/VistaLLM/.

Sanjoy Chowdhury, Sayan Nag, Dinesh Manocha

The choice of input text prompt plays a critical role in the performance of Vision-Language Pretrained (VLP) models such as CLIP. We present APoLLo, a unified multi-modal approach that combines Adapter and Prompt learning for Vision-Language models. Our method is designed to substantially improve the generalization capabilities of VLP models when they are fine-tuned in a few-shot setting. We introduce trainable cross-attention-based adapter layers in conjunction with vision and language encoders to strengthen the alignment between the two modalities. We enforce consistency between the respective encoder branches (receiving augmented inputs) to prevent overfitting in downstream tasks. Our method is evaluated on three representative tasks: generalization to novel classes, cross-dataset evaluation, and unseen domain shifts. In practice, APoLLo achieves a relative gain up to 6.03% over MaPLe (SOTA) on novel classes for 10 diverse image recognition datasets.

Sanjoy Chowdhury, Mohamed Elmoghany, Yohan Abeysinghe et al.

Large multimodal models (LMMs) have shown remarkable progress in audio-visual understanding, yet they struggle with real-world scenarios that require complex reasoning across extensive video collections. Existing benchmarks for video question answering remain limited in scope, typically involving one clip per query, which falls short of representing the challenges of large-scale, audio-visual retrieval and reasoning encountered in practical applications. To bridge this gap, we introduce a novel task named AV-HaystacksQA, where the goal is to identify salient segments across different videos in response to a query and link them together to generate the most informative answer. To this end, we present AVHaystacks, an audio-visual benchmark comprising 3100 annotated QA pairs designed to assess the capabilities of LMMs in multi-video retrieval and temporal grounding task. Additionally, we propose a model-agnostic, multi-agent framework MAGNET to address this challenge, achieving up to 89% and 65% relative improvements over baseline methods on BLEU@4 and GPT evaluation scores in QA task on our proposed AVHaystacks. To enable robust evaluation of multi-video retrieval and temporal grounding for optimal response generation, we introduce two new metrics, STEM, which captures alignment errors between a ground truth and a predicted step sequence and MTGS, to facilitate balanced and interpretable evaluation of segment-level grounding performance. Project: https://schowdhury671.github.io/magnet_project/

Siminfar Samakoush Galougah, Rishie Raj, Sanjoy Chowdhury et al.

Current audio-visual (AV) benchmarks focus on final answer accuracy, overlooking the underlying reasoning process. This makes it difficult to distinguish genuine comprehension from correct answers derived through flawed reasoning or hallucinations. To address this, we introduce AURA (Audio-visual Understanding and Reasoning Assessment), a benchmark for evaluating the cross-modal reasoning capabilities of Audio-Visual Large Language Models (AV-LLMs) and Omni-modal Language Models (OLMs). AURA includes questions across six challenging cognitive domains, such as causality, timbre and pitch, tempo and AV synchronization, unanswerability, implicit distractions, and skill profiling, explicitly designed to be unanswerable from a single modality. This forces models to construct a valid logical path grounded in both audio and video, setting AURA apart from AV datasets that allow uni-modal shortcuts. To assess reasoning traces, we propose a novel metric, AuraScore, which addresses the lack of robust tools for evaluating reasoning fidelity. It decomposes reasoning into two aspects: (i) Factual Consistency - whether reasoning is grounded in perceptual evidence, and (ii) Core Inference - the logical validity of each reasoning step. Evaluations of SOTA models on AURA reveal a critical reasoning gap: although models achieve high accuracy (up to 92% on some tasks), their Factual Consistency and Core Inference scores fall below 45%. This discrepancy highlights that models often arrive at correct answers through flawed logic, underscoring the need for our benchmark and paving the way for more robust multimodal evaluation.

Sanjoy Chowdhury, Subrata Biswas, Sayan Nag et al.

Modern perception models, particularly those designed for multisensory egocentric tasks, have achieved remarkable performance but often come with substantial computational costs. These high demands pose challenges for real-world deployment, especially in resource-constrained environments. In this paper, we introduce EgoAdapt, a framework that adaptively performs cross-modal distillation and policy learning to enable efficient inference across different egocentric perception tasks, including egocentric action recognition, active speaker localization, and behavior anticipation. Our proposed policy module is adaptable to task-specific action spaces, making it broadly applicable. Experimental results on three challenging egocentric datasets EPIC-Kitchens, EasyCom, and Aria Everyday Activities demonstrate that our method significantly enhances efficiency, reducing GMACs by up to 89.09%, parameters up to 82.02%, and energy up to 9.6x, while still on-par and in many cases outperforming, the performance of corresponding state-of-the-art models.

Sayan Nag, K J Joseph, Koustava Goswami et al.

Evaluating graphic designs involves assessing it from multiple facets like alignment, composition, aesthetics and color choices. Evaluating designs in a holistic way involves aggregating feedback from individual expert reviewers. Towards this, we propose an Agentic Design Review System (AgenticDRS), where multiple agents collaboratively analyze a design, orchestrated by a meta-agent. A novel in-context exemplar selection approach based on graph matching and a unique prompt expansion method plays central role towards making each agent design aware. Towards evaluating this framework, we propose DRS-BENCH benchmark. Thorough experimental evaluation against state-of-the-art baselines adapted to the problem setup, backed-up with critical ablation experiments brings out the efficacy of Agentic-DRS in evaluating graphic designs and generating actionable feedback. We hope that this work will attract attention to this pragmatic, yet under-explored research direction.

Arman Zarei, Samyadeep Basu, Keivan Rezaei et al.

Understanding how knowledge is distributed across the layers of generative models is crucial for improving interpretability, controllability, and adaptation. While prior work has explored knowledge localization in UNet-based architectures, Diffusion Transformer (DiT)-based models remain underexplored in this context. In this paper, we propose a model- and knowledge-agnostic method to localize where specific types of knowledge are encoded within the DiT blocks. We evaluate our method on state-of-the-art DiT-based models, including PixArt-alpha, FLUX, and SANA, across six diverse knowledge categories. We show that the identified blocks are both interpretable and causally linked to the expression of knowledge in generated outputs. Building on these insights, we apply our localization framework to two key applications: model personalization and knowledge unlearning. In both settings, our localized fine-tuning approach enables efficient and targeted updates, reducing computational cost, improving task-specific performance, and better preserving general model behavior with minimal interference to unrelated or surrounding content. Overall, our findings offer new insights into the internal structure of DiTs and introduce a practical pathway for more interpretable, efficient, and controllable model editing.

Sanjoy Chowdhury, Sayan Nag, K J Joseph et al.

Music is a universal language that can communicate emotions and feelings. It forms an essential part of the whole spectrum of creative media, ranging from movies to social media posts. Machine learning models that can synthesize music are predominantly conditioned on textual descriptions of it. Inspired by how musicians compose music not just from a movie script, but also through visualizations, we propose MeLFusion, a model that can effectively use cues from a textual description and the corresponding image to synthesize music. MeLFusion is a text-to-music diffusion model with a novel "visual synapse", which effectively infuses the semantics from the visual modality into the generated music. To facilitate research in this area, we introduce a new dataset MeLBench, and propose a new evaluation metric IMSM. Our exhaustive experimental evaluation suggests that adding visual information to the music synthesis pipeline significantly improves the quality of generated music, measured both objectively and subjectively, with a relative gain of up to 67.98% on the FAD score. We hope that our work will gather attention to this pragmatic, yet relatively under-explored research area.

Mayukh Bhattacharyya, Sayan Nag, Udita Ghosh

Air pollution poses a serious threat to sustainable environmental conditions in the 21st century. Its importance in determining the health and living standards in urban settings is only expected to increase with time. Various factors ranging from artificial emissions to natural phenomena are known to be primary causal agents or influencers behind rising air pollution levels. However, the lack of large scale data involving the major artificial and natural factors has hindered the research on the causes and relations governing the variability of the different air pollutants. Through this work, we introduce a large scale city-wise dataset for exploring the relationships among these agents over a long period of time. We also introduce a transformer based model - cosSquareFormer, for the problem of pollutant level estimation and forecasting. Our model outperforms most of the benchmark models for this task. We also analyze and explore the dataset through our model and other methodologies to bring out important inferences which enable us to understand the dynamics of the causal agents at a deeper level. Through our paper, we seek to provide a great set of foundations for further research into this domain that will demand critical attention of ours in the near future.

Sayan Nag, Mayukh Bhattacharyya

Activation functions play a pivotal role in determining the training dynamics and neural network performance. The widely adopted activation function ReLU despite being simple and effective has few disadvantages including the Dying ReLU problem. In order to tackle such problems, we propose a novel activation function called Serf which is self-regularized and nonmonotonic in nature. Like Mish, Serf also belongs to the Swish family of functions. Based on several experiments on computer vision (image classification and object detection) and natural language processing (machine translation, sentiment classification and multimodal entailment) tasks with different state-of-the-art architectures, it is observed that Serf vastly outperforms ReLU (baseline) and other activation functions including both Swish and Mish, with a markedly bigger margin on deeper architectures. Ablation studies further demonstrate that Serf based architectures perform better than those of Swish and Mish in varying scenarios, validating the effectiveness and compatibility of Serf with varying depth, complexity, optimizers, learning rates, batch sizes, initializers and dropout rates. Finally, we investigate the mathematical relation between Swish and Serf, thereby showing the impact of preconditioner function ingrained in the first derivative of Serf which provides a regularization effect making gradients smoother and optimization faster.

Sayan Nag

Self-supervised learning and pre-training strategieshave developed over the last few years especiallyfor Convolutional Neural Networks (CNNs). Re-cently application of such methods can also be no-ticed for Graph Neural Networks (GNNs) . In thispaper, we have used a graph based self-supervisedlearning strategy with different loss functions (Bar-low Twins[Zbontaret al., 2021], HSIC[Tsaiet al.,2021], VICReg[Bardeset al., 2021]) which haveshown promising results when applied with CNNspreviously. We have also proposed a hybrid lossfunction combining the advantages of VICReg andHSIC and called it as VICRegHSIC. The perfor-mance of these aforementioned methods have beencompared when applied to 7 different datasets suchas MUTAG, PROTEINS, IMDB-Binary, etc. Ex-periments showed that our hybrid loss function per-formed better than the remaining ones in 4 out of7 cases. Moreover, the impact of different batchsizes, projector dimensions and data augmentationstrategies have also been explored.

Sayan Nag, Uddalok Sarkar, Shankha Sanyal et al.

It is already known that both auditory and visual stimulus is able to convey emotions in human mind to different extent. The strength or intensity of the emotional arousal vary depending on the type of stimulus chosen. In this study, we try to investigate the emotional arousal in a cross-modal scenario involving both auditory and visual stimulus while studying their source characteristics. A robust fractal analytic technique called Detrended Fluctuation Analysis (DFA) and its 2D analogue has been used to characterize three (3) standardized audio and video signals quantifying their scaling exponent corresponding to positive and negative valence. It was found that there is significant difference in scaling exponents corresponding to the two different modalities. Detrended Cross Correlation Analysis (DCCA) has also been applied to decipher degree of cross-correlation among the individual audio and visual stimulus. This is the first of its kind study which proposes a novel algorithm with which emotional arousal can be classified in cross-modal scenario using only the source audio and visual signals while also attempting a correlation between them.

Uddalok Sarkar, Sayan Nag, Chirayata Bhattacharya et al.

At present emotion extraction from speech is a very important issue due to its diverse applications. Hence, it becomes absolutely necessary to obtain models that take into consideration the speaking styles of a person, vocal tract information, timbral qualities and other congenital information regarding his voice. Our speech production system is a nonlinear system like most other real world systems. Hence the need arises for modelling our speech information using nonlinear techniques. In this work we have modelled our articulation system using nonlinear multifractal analysis. The multifractal spectral width and scaling exponents reveals essentially the complexity associated with the speech signals taken. The multifractal spectrums are well distinguishable the in low fluctuation region in case of different emotions. The source characteristics have been quantified with the help of different non-linear models like Multi-Fractal Detrended Fluctuation Analysis, Wavelet Transform Modulus Maxima. The Results obtained from this study gives a very good result in emotion clustering.

Uddalok Sarkar, Sayan Nag, Medha Basu et al.

Music is often considered as the language of emotions. It has long been known to elicit emotions in human being and thus categorizing music based on the type of emotions they induce in human being is a very intriguing topic of research. When the task comes to classify emotions elicited by Indian Classical Music (ICM), it becomes much more challenging because of the inherent ambiguity associated with ICM. The fact that a single musical performance can evoke a variety of emotional response in the audience is implicit to the nature of ICM renditions. With the rapid advancements in the field of Deep Learning, this Music Emotion Recognition (MER) task is becoming more and more relevant and robust, hence can be applied to one of the most challenging test case i.e. classifying emotions elicited from ICM. In this paper we present a new dataset called JUMusEmoDB which presently has 400 audio clips (30 seconds each) where 200 clips correspond to happy emotions and the remaining 200 clips correspond to sad emotion. For supervised classification purposes, we have used 4 existing deep Convolutional Neural Network (CNN) based architectures (resnet18, mobilenet v2.0, squeezenet v1.0 and vgg16) on corresponding music spectrograms of the 2000 sub-clips (where every clip was segmented into 5 sub-clips of about 5 seconds each) which contain both time as well as frequency domain information. The initial results are quite inspiring, and we look forward to setting the baseline values for the dataset using this architecture. This type of CNN based classification algorithm using a rich corpus of Indian Classical Music is unique even in the global perspective and can be replicated in other modalities of music also. This dataset is still under development and we plan to include more data containing other emotional features as well. We plan to make the dataset publicly available soon.

Sayan Nag

Autoencoders are a class of artificial neural networks which have gained a lot of attention in the recent past. Using the encoder block of an autoencoder the input image can be compressed into a meaningful representation. Then a decoder is employed to reconstruct the compressed representation back to a version which looks like the input image. It has plenty of applications in the field of data compression and denoising. Another version of Autoencoders (AE) exist, called Variational AE (VAE) which acts as a generative model like GAN. Recently, an optimizer was introduced which is known as lookahead optimizer which significantly enhances the performances of Adam as well as SGD. In this paper, we implement Convolutional Autoencoders (CAE) and Convolutional Variational Autoencoders (CVAE) with lookahead optimizer (with Adam) and compare them with the Adam (only) optimizer counterparts. For this purpose, we have used a movie dataset comprising of natural images for the former case and CIFAR100 for the latter case. We show that lookahead optimizer (with Adam) improves the performance of CAEs for reconstruction of natural images.

Uddalok Sarkar, Soumyadeep Pal, Sayan Nag et al.

At present Automatic Speaker Recognition system is a very important issue due to its diverse applications. Hence, it becomes absolutely necessary to obtain models that take into consideration the speaking style of a person, vocal tract information, timbral qualities of his voice and other congenital information regarding his voice. The study of Bengali speech recognition and speaker identification is scarce in the literature. Hence the need arises for involving Bengali subjects in modelling our speaker identification engine. In this work, we have extracted some acoustic features of speech using non linear multifractal analysis. The Multifractal Detrended Fluctuation Analysis reveals essentially the complexity associated with the speech signals taken. The source characteristics have been quantified with the help of different techniques like Correlation Matrix, skewness of MFDFA spectrum etc. The Results obtained from this study gives a good recognition rate for Bengali Speakers.

Chirayata Bhattacharyya, Sourya Sengupta, Sayan Nag et al.

A recitation is a way of combining the words together so that they have a sense of rhythm and thus an emotional content is imbibed within. In this study we envisaged to answer these questions in a scientific manner taking into consideration 5 (five) well known Bengali recitations of different poets conveying a variety of moods ranging from joy to sorrow. The clips were recited as well as read (in the form of flat speech without any rhythm) by the same person to avoid any perceptual difference arising out of timbre variation. Next, the emotional content from the 5 recitations were standardized with the help of listening test conducted on a pool of 50 participants. The recitations as well as the speech were analyzed with the help of a latest non linear technique called Detrended Fluctuation Analysis (DFA) that gives a scaling exponent α, which is essentially the measure of long range correlations present in the signal. Similar pieces (the parts which have the exact lyrical content in speech as well as in the recital) were extracted from the complete signal and analyzed with the help of DFA technique. Our analysis shows that the scaling exponent for all parts of recitation were much higher in general as compared to their counterparts in speech. We have also established a critical value from our analysis, above which a mere speech may become a recitation. The case may be similar to the conventional phase transition, wherein the measurement of external condition at which the transformation occurs (generally temperature) is called phase transition. Further, we have also categorized the 5 recitations on the basis of their emotional content with the help of the same DFA technique. Analysis with a greater variety of recitations is being carried out to yield more interesting results.

Mayukh Bhattacharyya, Sayan Nag

Image Retrieval grows to be an integral part of fashion e-commerce ecosystem as it keeps expanding in multitudes. Other than the retrieval of visually similar items, the retrieval of visually compatible or complementary items is also an important aspect of it. Normal Siamese Networks tend to work well on complementary items retrieval. But it fails to identify low level style features which make items compatible in human eyes. These low level style features are captured to a large extent in techniques used in neural style transfer. This paper proposes a mechanism of utilising those methods in this retrieval task and capturing the low level style features through a hybrid siamese network coupled with a hybrid loss. The experimental results indicate that the proposed method outperforms traditional siamese networks in retrieval tasks for complementary items.

Sayan Nag, Shankha Sanyal, Archi Banerjee et al.

Can we hear the sound of our brain? Is there any technique which can enable us to hear the neuro-electrical impulses originating from the different lobes of brain? The answer to all these questions is YES. In this paper we present a novel method with which we can sonify the Electroencephalogram (EEG) data recorded in rest state as well as under the influence of a simplest acoustical stimuli - a tanpura drone. The tanpura drone has a very simple yet very complex acoustic features, which is generally used for creation of an ambiance during a musical performance. Hence, for this pilot project we chose to study the correlation between a simple acoustic stimuli (tanpura drone) and sonified EEG data. Till date, there have been no study which deals with the direct correlation between a bio-signal and its acoustic counterpart and how that correlation varies under the influence of different types of stimuli. This is the first of its kind study which bridges this gap and looks for a direct correlation between music signal and EEG data using a robust mathematical microscope called Multifractal Detrended Cross Correlation Analysis (MFDXA). For this, we took EEG data of 10 participants in 2 min 'rest state' (i.e. with white noise) and in 2 min 'tanpura drone' (musical stimulus) listening condition. Next, the EEG signals from different electrodes were sonified and MFDXA technique was used to assess the degree of correlation (or the cross correlation coefficient) between tanpura signal and EEG signals. The variation of γx for different lobes during the course of the experiment also provides major interesting new information. Only music stimuli has the ability to engage several areas of the brain significantly unlike other stimuli (which engages specific domains only).

Sayan Nag

Image Registration is the process of aligning two or more images of the same scene with reference to a particular image. The images are captured from various sensors at different times and at multiple view-points. Thus to get a better picture of any change of a scene or object over a considerable period of time image registration is important. Image registration finds application in medical sciences, remote sensing and in computer vision. This paper presents a detailed review of several approaches which are classified accordingly along with their contributions and drawbacks. The main steps of an image registration procedure are also discussed. Different performance measures are presented that determine the registration quality and accuracy. The scope for the future research are presented as well.

Uddalok Sarkar, Sayan Nag

In this paper a Metaheuristic approach for solving the N-Queens Problem is introduced to find the best possible solution in a reasonable amount of time. Genetic Algorithm is used with a novel fitness function as the Metaheuristic. The aim of N-Queens Problem is to place N queens on an N x N chessboard, in a way so that no queen is in conflict with the others. Chromosome representation and genetic operations like Mutation and Crossover are described in detail. Results show that this approach yields promising and satisfactory results in less time compared to that obtained from the previous approaches for several large values of N.

Sayan Nag

Vector Quantization, VQ is a popular image compression technique with a simple decoding architecture and high compression ratio. Codebook designing is the most essential part in Vector Quantization. LindeBuzoGray, LBG is a traditional method of generation of VQ Codebook which results in lower PSNR value. A Codebook affects the quality of image compression, so the choice of an appropriate codebook is a must. Several optimization techniques have been proposed for global codebook generation to enhance the quality of image compression. In this paper, a novel algorithm called IDE-LBG is proposed which uses Improved Differential Evolution Algorithm coupled with LBG for generating optimum VQ Codebooks. The proposed IDE works better than the traditional DE with modifications in the scaling factor and the boundary control mechanism. The IDE generates better solutions by efficient exploration and exploitation of the search space. Then the best optimal solution obtained by the IDE is provided as the initial Codebook for the LBG. This approach produces an efficient Codebook with less computational time and the consequences include excellent PSNR values and superior quality reconstructed images. It is observed that the proposed IDE-LBG find better VQ Codebooks as compared to IPSO-LBG, BA-LBG and FA-LBG.

Sayan Nag

One of the most straightforward, direct and efficient approaches to Image Segmentation is Image Thresholding. Multi-level Image Thresholding is an essential viewpoint in many image processing and Pattern Recognition based real-time applications which can effectively and efficiently classify the pixels into various groups denoting multiple regions in an Image. Thresholding based Image Segmentation using fuzzy entropy combined with intelligent optimization approaches are commonly used direct methods to properly identify the thresholds so that they can be used to segment an Image accurately. In this paper a novel approach for multi-level image thresholding is proposed using Type II Fuzzy sets combined with Adaptive Plant Propagation Algorithm (APPA). Obtaining the optimal thresholds for an image by maximizing the entropy is extremely tedious and time consuming with increase in the number of thresholds. Hence, Adaptive Plant Propagation Algorithm (APPA), a memetic algorithm based on plant intelligence, is used for fast and efficient selection of optimal thresholds. This fact is reasonably justified by comparing the accuracy of the outcomes and computational time consumed by other modern state-of-the-art algorithms such as Particle Swarm Optimization (PSO), Gravitational Search Algorithm (GSA) and Genetic Algorithm (GA).

Sayan Nag

Optimization problems in design engineering are complex by nature, often because of the involvement of critical objective functions accompanied by a number of rigid constraints associated with the products involved. One such problem is Economic Load Dispatch (ED) problem which focuses on the optimization of the fuel cost while satisfying some system constraints. Classical optimization algorithms are not sufficient and also inefficient for the ED problem involving highly nonlinear, and non-convex functions both in the objective and in the constraints. This led to the development of metaheuristic optimization approaches which can solve the ED problem almost efficiently. This paper presents a novel robust plant intelligence based Adaptive Plant Propagation Algorithm (APPA) which is used to solve the classical ED problem. The application of the proposed method to the 3-generator and 6-generator systems shows the efficiency and robustness of the proposed algorithm. A comparative study with another state-of-the-art algorithm (APSO) demonstrates the quality of the solution achieved by the proposed method along with the convergence characteristics of the proposed approach.

Sayan Nag, Sayan Biswas, Sourya Sengupta et al.

The term jitter and shimmer has long been used in the domain of speech and acoustic signal analysis as a parameter for speaker identification and other prosodic features. In this study, we look forward to use the same parameters in neural domain to identify and categorize emotional cues in different musical clips. For this, we chose two ragas of Hindustani music which are conventionally known to portray contrast emotions and EEG study was conducted on 5 participants who were made to listen to 3 min clip of these two ragas with sufficient resting period in between. The neural jitter and shimmer components were evaluated for each experimental condition. The results reveal interesting information regarding domain specific arousal of human brain in response to musical stimuli and also regarding trait characteristics of an individual. This novel study can have far reaching conclusions when it comes to modeling of emotional appraisal. The results and implications are discussed in detail.

Shankha Sanyal, Archi Banerjee, Souparno Roy et al.

The term gestalt has been widely used in the field of psychology which defined the perception of human mind to group any object not in part but as a unified whole. Music in general is polytonic i.e. a combination of a number of pure tones (frequencies) mixed together in a manner that sounds harmonius. The study of human brain response due to different frequency groups of acoustic signal can give us an excellent insight regarding the neural and functional architecture of brain functions. In this work we have tried to analyze the effect of different frequency bands of music on the various frequency rhythms of human brain obtained from EEG data of 5 participants. Four (4) widely popular Rabindrasangeet clips were subjected to Wavelet Transform method for extracting five resonant frequency bands from the original music signal. These resonant frequency bands were presented to the subjects as auditory stimulus and EEG signals recorded simultaneously in 19 different locations of the brain. The recorded EEG signals were noise cleaned and subjected to Multifractal Detrended Fluctuation Analysis (MFDFA) technique on the alpha, theta and gamma frequency range. Thus, we obtained the complexity values (in the form of multifractal spectral width) in alpha, theta and gamma EEG rhythms corresponding to different frequency bands of music. We obtain frequency specific arousal based response in different lobes of brain as well as in specific EEG bands corresponding to musical stimuli. This revelation can be of immense importance when it comes to the field of cognitive music therapy.

Shankha Sanyal, Archi Banerjee, Souparno Roy et al.

Since the start of Indian cinema, a number of films have been made where a particular song is based on a certain raga. These songs have been taking a major role in spreading the essence of classical music to the common people, who have no formal exposure to classical music. In this paper, we look to explore what are the particular features of a certain raga which make it understandable to common people and enrich the song to a great extent. For this, we chose two common ragas of Hindustani classical music, namely "Bhairav" and "Mian ki Malhar" which are known to have widespread application in popular film music. We have taken 3 minute clips of these two ragas from the renderings of two eminent maestros of Hindustani classical music. 3 min clips of ten (10) widely popular songs of Bollywood films were selected for analysis. These were analyzed with the help of a latest non linear analysis technique called Multifractal Detrended Cross correlation Analysis (MFDXA). With this technique, all parts of the Film music and the renderings from the eminent maestros are analyzed to find out a cross correlation coefficient (γx) which gives the degree of correlation between these two signals. We hypothesize that the parts which have the highest degree of cross correlation are the parts in which that particular raga is established in the song. Also the variation of cross correlation coefficient in the different parts of the two samples gives a measure of the modulation that is executed by the singer. Thus, in nutshell we try to study scientifically the amount of correlation that exists between the raga and the same raga being utilized in Film music. This will help in generating an automated algorithm through which a naïve listener will relish the flavor of a particular raga in a popular film song. The results are discussed in detail.