Mausoom Sarkar

Sukriti Verma, Ayush Chopra, Jayakumar Subramanian et al.

The two-time scale nature of SAC, which is an actor-critic algorithm, is characterised by the fact that the critic estimate has not converged for the actor at any given time, but since the critic learns faster than the actor, it ensures eventual consistency between the two. Various strategies have been introduced in literature to learn better gradient estimates to help achieve better convergence. Since gradient estimates depend upon the critic, we posit that improving the critic can provide a better gradient estimate for the actor at each time. Utilizing this, we propose Soft Actor Retrospective Critic (SARC), where we augment the SAC critic loss with another loss term - retrospective loss - leading to faster critic convergence and consequently, better policy gradient estimates for the actor. An existing implementation of SAC can be easily adapted to SARC with minimal modifications. Through extensive experimentation and analysis, we show that SARC provides consistent improvement over SAC on benchmark environments. We plan to open-source the code and all experiment data at: https://github.com/sukritiverma1996/SARC.

Rishabh Jain, Krishna Kumar Singh, Mayur Hemani et al.

The task of human reposing involves generating a realistic image of a person standing in an arbitrary conceivable pose. There are multiple difficulties in generating perceptually accurate images, and existing methods suffer from limitations in preserving texture, maintaining pattern coherence, respecting cloth boundaries, handling occlusions, manipulating skin generation, etc. These difficulties are further exacerbated by the fact that the possible space of pose orientation for humans is large and variable, the nature of clothing items is highly non-rigid, and the diversity in body shape differs largely among the population. To alleviate these difficulties and synthesize perceptually accurate images, we propose VGFlow. Our model uses a visibility-guided flow module to disentangle the flow into visible and invisible parts of the target for simultaneous texture preservation and style manipulation. Furthermore, to tackle distinct body shapes and avoid network artifacts, we also incorporate a self-supervised patch-wise "realness" loss to improve the output. VGFlow achieves state-of-the-art results as observed qualitatively and quantitatively on different image quality metrics (SSIM, LPIPS, FID).

Mausoom Sarkar, Nikitha SR, Mayur Hemani et al.

Face parsing is defined as the per-pixel labeling of images containing human faces. The labels are defined to identify key facial regions like eyes, lips, nose, hair, etc. In this work, we make use of the structural consistency of the human face to propose a lightweight face-parsing method using a Local Implicit Function network, FP-LIIF. We propose a simple architecture having a convolutional encoder and a pixel MLP decoder that uses 1/26th number of parameters compared to the state-of-the-art models and yet matches or outperforms state-of-the-art models on multiple datasets, like CelebAMask-HQ and LaPa. We do not use any pretraining, and compared to other works, our network can also generate segmentation at different resolutions without any changes in the input resolution. This work enables the use of facial segmentation on low-compute or low-bandwidth devices because of its higher FPS and smaller model size.

Silky Singh, Shripad Deshmukh, Mausoom Sarkar et al.

Learning object segmentation in image and video datasets without human supervision is a challenging problem. Humans easily identify moving salient objects in videos using the gestalt principle of common fate, which suggests that what moves together belongs together. Building upon this idea, we propose a self-supervised object discovery approach that leverages motion and appearance information to produce high-quality object segmentation masks. Specifically, we redesign the traditional graph cut on images to include motion information in a linear combination with appearance information to produce edge weights. Remarkably, this step produces object segmentation masks comparable to the current state-of-the-art on multiple benchmarks. To further improve performance, we bootstrap a segmentation network trained on these preliminary masks as pseudo-ground truths to learn from its own outputs via self-training. We demonstrate the effectiveness of our approach, named LOCATE, on multiple standard video object segmentation, image saliency detection, and object segmentation benchmarks, achieving results on par with and, in many cases surpassing state-of-the-art methods. We also demonstrate the transferability of our approach to novel domains through a qualitative study on in-the-wild images. Additionally, we present extensive ablation analysis to support our design choices and highlight the contribution of each component of our proposed method.

Varun Gopal, Rishabh Jain, Aradhya Mathur et al.

Graphic layouts serve as an important and engaging medium for visual communication across different channels. While recent layout generation models have demonstrated impressive capabilities, they frequently fail to align with nuanced human aesthetic judgment. Existing preference datasets and reward models trained on text-to-image generation do not generalize to layout evaluation, where the spatial arrangement of identical elements determines quality. To address this critical gap, we introduce DesignSense-10k, a large-scale dataset of 10,235 human-annotated preference pairs for graphic layout evaluation. We propose a five-stage curation pipeline that generates visually coherent layout transformations across diverse aspect ratios, using semantic grouping, layout prediction, filtering, clustering, and VLM-based refinement to produce high-quality comparison pairs. Human preferences are annotated using a 4-class scheme (left, right, both good, both bad) to capture subjective ambiguity. Leveraging this dataset, we train DesignSense, a vision-language model-based classifier that substantially outperforms existing open-source and proprietary models across comprehensive evaluation metrics (54.6% improvement in Macro F1 over the strongest proprietary baseline). Our analysis shows that frontier VLMs remain unreliable overall and fail catastrophically on the full four-class task, underscoring the need for specialized, preference-aware models. Beyond the dataset, our reward model DesignSense yields tangible downstream gains in layout generation. Using our judge during RL based training improves generator win rate by about 3%, while inference-time scaling, which involves generating multiple candidates and selecting the best one, provides a 3.6% improvement. These results highlight the practical impact of specialized, layout-aware preference modeling on real-world layout generation quality.

Rishabh Jain, Mayur Hemani, Duygu Ceylan et al.

Numerous pose-guided human editing methods have been explored by the vision community due to their extensive practical applications. However, most of these methods still use an image-to-image formulation in which a single image is given as input to produce an edited image as output. This objective becomes ill-defined in cases when the target pose differs significantly from the input pose. Existing methods then resort to in-painting or style transfer to handle occlusions and preserve content. In this paper, we explore the utilization of multiple views to minimize the issue of missing information and generate an accurate representation of the underlying human model. To fuse knowledge from multiple viewpoints, we design a multi-view fusion network that takes the pose key points and texture from multiple source images and generates an explainable per-pixel appearance retrieval map. Thereafter, the encodings from a separate network (trained on a single-view human reposing task) are merged in the latent space. This enables us to generate accurate, precise, and visually coherent images for different editing tasks. We show the application of our network on two newly proposed tasks - Multi-view human reposing and Mix&Match Human Image generation. Additionally, we study the limitations of single-view editing and scenarios in which multi-view provides a better alternative.

Silky Singh, Shripad Deshmukh, Mausoom Sarkar et al.

Segmentation of objects in a video is challenging due to the nuances such as motion blurring, parallax, occlusions, changes in illumination, etc. Instead of addressing these nuances separately, we focus on building a generalizable solution that avoids overfitting to the individual intricacies. Such a solution would also help us save enormous resources involved in human annotation of video corpora. To solve Video Object Segmentation (VOS) in an unsupervised setting, we propose a new pipeline (FODVid) based on the idea of guiding segmentation outputs using flow-guided graph-cut and temporal consistency. Basically, we design a segmentation model incorporating intra-frame appearance and flow similarities, and inter-frame temporal continuation of the objects under consideration. We perform an extensive experimental analysis of our straightforward methodology on the standard DAVIS16 video benchmark. Though simple, our approach produces results comparable (within a range of ~2 mIoU) to the existing top approaches in unsupervised VOS. The simplicity and effectiveness of our technique opens up new avenues for research in the video domain.

Abhinav Java, Shripad Deshmukh, Milan Aggarwal et al.

Active consumption of digital documents has yielded scope for research in various applications, including search. Traditionally, searching within a document has been cast as a text matching problem ignoring the rich layout and visual cues commonly present in structured documents, forms, etc. To that end, we ask a mostly unexplored question: "Can we search for other similar snippets present in a target document page given a single query instance of a document snippet?". We propose MONOMER to solve this as a one-shot snippet detection task. MONOMER fuses context from visual, textual, and spatial modalities of snippets and documents to find query snippet in target documents. We conduct extensive ablations and experiments showing MONOMER outperforms several baselines from one-shot object detection (BHRL), template matching, and document understanding (LayoutLMv3). Due to the scarcity of relevant data for the task at hand, we train MONOMER on programmatically generated data having many visually similar query snippets and target document pairs from two datasets - Flamingo Forms and PubLayNet. We also do a human study to validate the generated data.

Tarun Ram Menta, Surgan Jandial, Akash Patil et al.

As transfer learning techniques are increasingly used to transfer knowledge from the source model to the target task, it becomes important to quantify which source models are suitable for a given target task without performing computationally expensive fine tuning. In this work, we propose HASTE (HArd Subset TransfErability), a new strategy to estimate the transferability of a source model to a particular target task using only a harder subset of target data. By leveraging the internal and output representations of model, we introduce two techniques, one class agnostic and another class specific, to identify harder subsets and show that HASTE can be used with any existing transferability metric to improve their reliability. We further analyze the relation between HASTE and the optimal average log likelihood as well as negative conditional entropy and empirically validate our theoretical bounds. Our experimental results across multiple source model architectures, target datasets, and transfer learning tasks show that HASTE modified metrics are consistently better or on par with the state of the art transferability metrics.

Neeraj Anand, Rishabh Jain, Sohan Patnaik et al.

There is a growing demand for mobile user interface (UI) automation, driven by its broad applications across industries. With the advent of visual language models (VLMs), GUI automation has progressed from generating text-based instructions for humans to autonomously executing tasks, thus optimizing automation workflows. Recent approaches leverage VLMs for this problem due to their ability to 1) process on-screen content directly, 2) remain independent of device-specific APIs by utilizing human actions (e.g., clicks, typing), and 3) apply real-world contextual knowledge for task understanding. However, these models often have trouble accurately identifying widgets and determining actions due to limited spatial information in vision encoder features. Additionally, top-performing models are often large, requiring extensive training and resulting in inference delays. In this work, we introduce AFRAgent, an instruct-BLIP-based multimodal architecture that achieves superior performance in GUI automation while being less than one-fourth the size of its nearest competitor. To enhance image embeddings in the large language model (LLM) pipeline, we propose an adaptive feature renormalization-based (a token-level affine transformation) technique that effectively enriches low-resolution image embeddings and fuses high-resolution details. We evaluate AFRAgent on Meta-GUI and AITW benchmarks, establishing a new state-of-the-art baseline for smartphone automation.

Akilesh B, Abhishek Sinha, Mausoom Sarkar et al.

In this work, we focus on the problem of grounding language by training an agent to follow a set of natural language instructions and navigate to a target object in an environment. The agent receives visual information through raw pixels and a natural language instruction telling what task needs to be achieved and is trained in an end-to-end way. We develop an attention mechanism for multi-modal fusion of visual and textual modalities that allows the agent to learn to complete the task and achieve language grounding. Our experimental results show that our attention mechanism outperforms the existing multi-modal fusion mechanisms proposed for both 2D and 3D environments in order to solve the above-mentioned task in terms of both speed and success rate. We show that the learnt textual representations are semantically meaningful as they follow vector arithmetic in the embedding space. The effectiveness of our attention approach over the contemporary fusion mechanisms is also highlighted from the textual embeddings learnt by the different approaches. We also show that our model generalizes effectively to unseen scenarios and exhibit zero-shot generalization capabilities both in 2D and 3D environments. The code for our 2D environment as well as the models that we developed for both 2D and 3D are available at https://github.com/rl-lang-grounding/rl-lang-ground.

Sohan Patnaik, Rishabh Jain, Balaji Krishnamurthy et al.

Visual layouts are essential in graphic design fields such as advertising, posters, and web interfaces. The application of generative models for content-aware layout generation has recently gained traction. However, these models fail to understand the contextual aesthetic requirements of layout design and do not align with human-like preferences, primarily treating it as a prediction task without considering the final rendered output. To overcome these problems, we offer Aesthetic-Aware Preference Alignment(AAPA), a novel technique to train a Multi-modal Large Language Model (MLLM) for layout prediction that uses MLLM's aesthetic preferences for Direct Preference Optimization over graphic layouts. We propose a data filtering protocol utilizing our layout-quality heuristics for AAPA to ensure training happens on high-quality layouts. Additionally, we introduce a novel evaluation metric that uses another MLLM to compute the win rate of the generated layout against the ground-truth layout based on aesthetics criteria. We also demonstrate the applicability of AAPA for MLLMs of varying scales (1B to 8B parameters) and LLM families (Qwen, Phi, InternLM). By conducting thorough qualitative and quantitative analyses, we verify the efficacy of our approach on two challenging benchmarks - Crello and Webui, showcasing 17%, and 16 improvement over current State-of-The-Art methods, thereby highlighting the potential of MLLMs in aesthetic-aware layout generation.

Nikitha SR, Aradhya Neeraj Mathur, Tarun Ram Menta et al.

The integration of high-resolution image features in modern multimodal large language models has demonstrated significant improvements in fine-grained visual understanding tasks, achieving high performance across multiple benchmarks. Since these features are obtained from large image encoders like ViT, they come with a significant increase in computational costs due to multiple calls to these encoders. In this work, we first develop an intuition for feature upsampling as a natural extension of high-resolution feature generation. Through extensive experiments and ablations, we demonstrate how a shallow feature enricher can achieve competitive results with tremendous reductions in training and inference time as well as computational cost, with upto 1.5x saving in FLOPs.

Sarthak Mehrotra, Rishabh Jain, Mayur Hemani et al.

Reposing objects in images has a myriad of applications, especially for e-commerce where several variants of product images need to be produced quickly. In this work, we leverage the recent advances in unsupervised keypoint correspondence detection between different object images of the same class to propose an end-to-end framework for generic object reposing. Our method, EOPose, takes a target pose-guidance image as input and uses its keypoint correspondence with the source object image to warp and re-render the latter into the target pose using a novel three-step approach. Unlike generative approaches, our method also preserves the fine-grained details of the object such as its exact colors, textures, and brand marks. We also prepare a new dataset of paired objects based on the Objaverse dataset to train and test our network. EOPose produces high-quality reposing output as evidenced by different image quality metrics (PSNR, SSIM and FID). Besides a description of the method and the dataset, the paper also includes detailed ablation and user studies to indicate the efficacy of the proposed method

Nikitha SR, Tarun Ram Menta, Mausoom Sarkar

The advent of multimodal learning has brought a significant improvement in document AI. Documents are now treated as multimodal entities, incorporating both textual and visual information for downstream analysis. However, works in this space are often focused on the textual aspect, using the visual space as auxiliary information. While some works have explored pure vision based techniques for document image understanding, they require OCR identified text as input during inference, or do not align with text in their learning procedure. Therefore, we present a novel image-text alignment technique specially designed for leveraging the textual information in document images to improve performance on visual tasks. Our document encoder model DoPTA - trained with this technique demonstrates strong performance on a wide range of document image understanding tasks, without requiring OCR during inference. Combined with an auxiliary reconstruction objective, DoPTA consistently outperforms larger models, while using significantly lesser pre-training compute. DoPTA also sets new state-of-the art results on D4LA, and FUNSD, two challenging document visual analysis benchmarks.

Sumedh A Sontakke, Sumegh Roychowdhury, Mausoom Sarkar et al.

Humans excel at learning long-horizon tasks from demonstrations augmented with textual commentary, as evidenced by the burgeoning popularity of tutorial videos online. Intuitively, this capability can be separated into 2 distinct subtasks - first, dividing a long-horizon demonstration sequence into semantically meaningful events; second, adapting such events into meaningful behaviors in one's own environment. Here, we present Video2Skill (V2S), which attempts to extend this capability to artificial agents by allowing a robot arm to learn from human cooking videos. We first use sequence-to-sequence Auto-Encoder style architectures to learn a temporal latent space for events in long-horizon demonstrations. We then transfer these representations to the robotic target domain, using a small amount of offline and unrelated interaction data (sequences of state-action pairs of the robot arm controlled by an expert) to adapt these events into actionable representations, i.e., skills. Through experiments, we demonstrate that our approach results in self-supervised analogy learning, where the agent learns to draw analogies between motions in human demonstration data and behaviors in the robotic environment. We also demonstrate the efficacy of our approach on model learning - demonstrating how Video2Skill utilizes prior knowledge from human demonstration to outperform traditional model learning of long-horizon dynamics. Finally, we demonstrate the utility of our approach for non-tabula rasa decision-making, i.e, utilizing video demonstration for zero-shot skill generation.

Milan Aggarwal, Hiresh Gupta, Mausoom Sarkar et al.

Document structure extraction has been a widely researched area for decades with recent works performing it as a semantic segmentation task over document images using fully-convolution networks. Such methods are limited by image resolution due to which they fail to disambiguate structures in dense regions which appear commonly in forms. To mitigate this, we propose Form2Seq, a novel sequence-to-sequence (Seq2Seq) inspired framework for structure extraction using text, with a specific focus on forms, which leverages relative spatial arrangement of structures. We discuss two tasks; 1) Classification of low-level constituent elements (TextBlock and empty fillable Widget) into ten types such as field captions, list items, and others; 2) Grouping lower-level elements into higher-order constructs, such as Text Fields, ChoiceFields and ChoiceGroups, used as information collection mechanism in forms. To achieve this, we arrange the constituent elements linearly in natural reading order, feed their spatial and textual representations to Seq2Seq framework, which sequentially outputs prediction of each element depending on the final task. We modify Seq2Seq for grouping task and discuss improvements obtained through cascaded end-to-end training of two tasks versus training in isolation. Experimental results show the effectiveness of our text-based approach achieving an accuracy of 90% on classification task and an F1 of 75.82, 86.01, 61.63 on groups discussed above respectively, outperforming segmentation baselines. Further we show our framework achieves state of the results for table structure recognition on ICDAR 2013 dataset.

Milan Aggarwal, Mausoom Sarkar, Hiresh Gupta et al.

Document structure extraction has been a widely researched area for decades. Recent work in this direction has been deep learning-based, mostly focusing on extracting structure using fully convolution NN through semantic segmentation. In this work, we present a novel multi-modal approach for form structure extraction. Given simple elements such as textruns and widgets, we extract higher-order structures such as TextBlocks, Text Fields, Choice Fields, and Choice Groups, which are essential for information collection in forms. To achieve this, we obtain a local image patch around each low-level element (reference) by identifying candidate elements closest to it. We process textual and spatial representation of candidates sequentially through a BiLSTM to obtain context-aware representations and fuse them with image patch features obtained by processing it through a CNN. Subsequently, the sequential decoder takes this fused feature vector to predict the association type between reference and candidates. These predicted associations are utilized to determine larger structures through connected components analysis. Experimental results show the effectiveness of our approach achieving a recall of 90.29%, 73.80%, 83.12%, and 52.72% for the above structures, respectively, outperforming semantic segmentation baselines significantly. We show the efficacy of our method through ablations, comparing it against using individual modalities. We also introduce our new rich human-annotated Forms Dataset.

Sumegh Roychowdhury, Sumedh A. Sontakke, Nikaash Puri et al.

Temporal event representations are an essential aspect of learning among humans. They allow for succinct encoding of the experiences we have through a variety of sensory inputs. Also, they are believed to be arranged hierarchically, allowing for an efficient representation of complex long-horizon experiences. Additionally, these representations are acquired in a self-supervised manner. Analogously, here we propose a model that learns temporal representations from long-horizon visual demonstration data and associated textual descriptions, without explicit temporal supervision. Our method produces a hierarchy of representations that align more closely with ground-truth human-annotated events (+15.3) than state-of-the-art unsupervised baselines. Our results are comparable to heavily-supervised baselines in complex visual domains such as Chess Openings, YouCook2 and TutorialVQA datasets. Finally, we perform ablation studies illustrating the robustness of our approach. We release our code and demo visualizations in the Supplementary Material.

Surgan Jandial, Pinkesh Badjatiya, Pranit Chawla et al.

The ability to efficiently search for images is essential for improving the user experiences across various products. Incorporating user feedback, via multi-modal inputs, to navigate visual search can help tailor retrieved results to specific user queries. We focus on the task of text-conditioned image retrieval that utilizes support text feedback alongside a reference image to retrieve images that concurrently satisfy constraints imposed by both inputs. The task is challenging since it requires learning composite image-text features by incorporating multiple cross-granular semantic edits from text feedback and then applying the same to visual features. To address this, we propose a novel framework SAC which resolves the above in two major steps: "where to see" (Semantic Feature Attention) and "how to change" (Semantic Feature Modification). We systematically show how our architecture streamlines the generation of text-aware image features by removing the need for various modules required by other state-of-art techniques. We present extensive quantitative, qualitative analysis, and ablation studies, to show that our architecture SAC outperforms existing techniques by achieving state-of-the-art performance on 3 benchmark datasets: FashionIQ, Shoes, and Birds-to-Words, while supporting natural language feedback of varying lengths.

Surgan Jandial, Ayush Chopra, Mausoom Sarkar et al.

Deep neural networks (DNNs) are powerful learning machines that have enabled breakthroughs in several domains. In this work, we introduce a new retrospective loss to improve the training of deep neural network models by utilizing the prior experience available in past model states during training. Minimizing the retrospective loss, along with the task-specific loss, pushes the parameter state at the current training step towards the optimal parameter state while pulling it away from the parameter state at a previous training step. Although a simple idea, we analyze the method as well as to conduct comprehensive sets of experiments across domains - images, speech, text, and graphs - to show that the proposed loss results in improved performance across input domains, tasks, and architectures.

Pinkesh Badjatiya, Mausoom Sarkar, Abhishek Sinha et al.

In social dilemma situations, individual rationality leads to sub-optimal group outcomes. Several human engagements can be modeled as a sequential (multi-step) social dilemmas. However, in contrast to humans, Deep Reinforcement Learning agents trained to optimize individual rewards in sequential social dilemmas converge to selfish, mutually harmful behavior. We introduce a status-quo loss (SQLoss) that encourages an agent to stick to the status quo, rather than repeatedly changing its policy. We show how agents trained with SQLoss evolve cooperative behavior in several social dilemma matrix games. To work with social dilemma games that have visual input, we propose GameDistill. GameDistill uses self-supervision and clustering to automatically extract cooperative and selfish policies from a social dilemma game. We combine GameDistill and SQLoss to show how agents evolve socially desirable cooperative behavior in the Coin Game.

Mausoom Sarkar, Milan Aggarwal, Arneh Jain et al.

Structure extraction from document images has been a long-standing research topic due to its high impact on a wide range of practical applications. In this paper, we share our findings on employing a hierarchical semantic segmentation network for this task of structure extraction. We propose a prior based deep hierarchical CNN network architecture that enables document structure extraction using very high resolution(1800 x 1000) images. We divide the document image into overlapping horizontal strips such that the network segments a strip and uses its prediction mask as prior for predicting the segmentation of the subsequent strip. We perform experiments establishing the effectiveness of our strip based network architecture through ablation methods and comparison with low-resolution variations. Further, to demonstrate our network's capabilities, we train it on only one type of documents (Forms) and achieve state-of-the-art results over other general document datasets. We introduce our new human-annotated forms dataset and show that our method significantly outperforms different segmentation baselines on this dataset in extracting hierarchical structures. Our method is currently being used in Adobe's AEM Forms for automated conversion of paper and PDF forms to modern HTML based forms.

Abhishek Sinha, Mausoom Sarkar, Aahitagni Mukherjee et al.

Neural Networks are function approximators that have achieved state-of-the-art accuracy in numerous machine learning tasks. In spite of their great success in terms of accuracy, their large training time makes it difficult to use them for various tasks. In this paper, we explore the idea of learning weight evolution pattern from a simple network for accelerating training of novel neural networks. We use a neural network to learn the training pattern from MNIST classification and utilize it to accelerate training of neural networks used for CIFAR-10 and ImageNet classification. Our method has a low memory footprint and is computationally efficient. This method can also be used with other optimizers to give faster convergence. The results indicate a general trend in the weight evolution during training of neural networks.