Papers by Sumit Shekhar

9 papers
CaRVE: Critiquing and Refining Visual Elaborations for Figurative Language Illustrations (2026.findings-acl)

Copied to clipboard

Challenge: Existing text-to-image frameworks for figurative illustration rely on proprietary models or human supervision to achieve adequate alignment.
Approach: They propose a critique-driven framework that uses VLM feedback to refine visual elaborations for figurative image generation.
Outcome: The proposed framework outperforms existing figurative image-to-text pipelines on human-supervised visual elaborations.
Looking Beyond the Pixels: Evaluating Visual Metaphor Understanding in VLMs (2025.findings-emnlp)

Copied to clipboard

Challenge: Visual metaphors are a complex vision–language phenomenon that requires both perceptual and conceptual reasoning to understand.
Approach: They introduce a visual metaphor dataset featuring 2177 synthetic and 350 human-annotated images and benchmark several SOTA VLMs on two tasks: Visual Metaphor Captioning (VMC) and Visual Metamorphosis VQA (VM-VQA).
Outcome: The proposed model outperforms standard few-shot baselines on visual metaphors and VM-VQA tasks.
STL-CQA: Structure-based Transformers with Localization and Encoding for Chart Question Answering (2020.emnlp-main)

Copied to clipboard

Challenge: Chart Question Answering (CQA) is a task of answering natural language questions about visualisations in the chart image.
Approach: They propose a method for Chart Question Answering which improves the question/answering through sequential elements localization, question encoding and then, a structural transformer-based learning approach.
Outcome: The proposed method outperforms state-of-the-art methods on various chart Q/A datasets while outperforming even human baseline.
“Let’s not Quote out of Context”: Unified Vision-Language Pretraining for Context Assisted Image Captioning (2023.acl-industry)

Copied to clipboard

Challenge: Large enterprises have several teams to create their content for the purpose of marketing, campaigning, or even maintaining a brand presence.
Approach: They propose a new unified Vision-Language (VL) model with a focus on context-assisted image captioning where the caption is generated based on both the image and its context.
Outcome: The proposed model achieves state-of-the-art with an improvement of up to 8.34 CIDEr score on the benchmark news image captioning datasets.
DynamicTOC: Persona-based Table of Contents for Consumption of Long Documents (2022.naacl-main)

Copied to clipboard

Challenge: Long documents are tedious to read through and can be authored by multiple entities . traditional document navigation is through a Table of Contents (ToC) but there is no way to highlight information relevant to different personas.
Approach: They propose a dynamic table of content-based navigator that highlights sections of interest . DYNAMICTOC is augmented with short questions to assist users in understanding underlying content .
Outcome: The proposed navigator highlights sections of interest in documents as per the aspects relevant to different personas. human and automatic evaluations suggest the efficacy of both end-to-end pipeline and different components.
Unveiling the Invisible: Captioning Videos with Metaphors (2024.findings-emnlp)

Copied to clipboard

Challenge: Recent studies have shown that Vision-Language models cannot understand visual metaphors in memes and adverts.
Approach: They propose a task to describe visual metaphors in videos using a manually created dataset and a new metric called Average Concept Distance to automatically evaluate creativity.
Outcome: The proposed system performs comparable to existing video language models on the proposed task and can be used for future research.
Open-World Factually Consistent Question Generation (2023.findings-acl)

Copied to clipboard

Challenge: Existing methods for question generation suffer from factual inconsistencies and incorrect entities and are not answerable from the input paragraph.
Approach: They propose a data processing technique based on de-lexicalization for consistent question generation across domains and a model that is generic across question-generation models.
Outcome: The proposed method produces entity-level factually consistent questions without significant impact on traditional metrics.
A Neural CRF-based Hierarchical Approach for Linear Text Segmentation (2023.findings-eacl)

Copied to clipboard

Challenge: Existing methods to segment unformatted text and transcripts explicitly train to predict segment boundaries, but they fail to provide a large annotated dataset.
Approach: They propose a method to generate hierarchical segmentation structures based on Wikipedia annotations by using a neural conditional random field.
Outcome: The proposed method outperforms or achieves competitive performance when compared to previous state-of-the-art algorithms.
Imposter: Text and Frequency Guidance for Subject Driven Action Personalization using Diffusion Models (2025.coling-main)

Copied to clipboard

Challenge: ImPoster is a novel algorithm for generating a target image of a ‘source’ subject performing a 'driving' action.
Approach: They propose an unsupervised approach that generates a target image of a ‘source’ subject performing a driving action from a single pair of inputs along with the text descriptions of the two images.
Outcome: The proposed algorithm is completely unsupervised and does not require access to additional annotations like keypoints or pose.

What is GenGO?

GenGO is an NLP powered publication search system. It currenctly indexes 30k+ papers from ACL Anthology, and implements multi-aspect summarization, semantic search, and more!

Information

About
Limitations