While Multi-modal Large Language Models (MLLMs) have shown impressive capabilities in document understanding tasks, their ability to locate and reason about fine-grained details within complex documents remains understudied. Consider searching a restaurant menu for a specific nutritional detail or identifying a disclaimer in a lengthy newspaper article — tasks that demand careful attention to small but significant details within a broader narrative, akin to Finding Needles in Images (NiM). To address this gap, we introduce NiM-Benchmark, a carefully curated benchmark spanning diverse real-world documents including newspapers, menus, and lecture images, specifically designed to evaluate MLLMs’ capability in these intricate tasks. Building on this, we further propose Spot-IT, a simple yet effective approach that enhances MLLMs capability through intelligent patch selection and Gaussian attention, motivated from how humans zoom and focus when searching documents. Our extensive experiments reveal both the capabilities and limitations of current MLLMs in handling fine-grained document understanding tasks, while demonstrating the effectiveness of our approach. Spot-IT achieves significant improvements over baseline methods, particularly in scenarios requiring precise detail extraction from complex layouts.

Enterprise systems are crucial for enhancing productivity and decision-making among employees and customers.Integrating LLM based systems into enterprise systems enables intelligent automation, personalized experiences, and efficient information retrieval, driving operational efficiency and strategic growth.However, developing and evaluating such systems is challenging due to the inherent complexity of enterprise environments, where data is fragmented across multiple sources and governed by sophisticated access controls.We present Enter-priseBench, a comprehensive benchmark that simulates enterprise settings, featuring 500 diverse tasks across software engineering, HR, finance, and administrative domains.Our benchmark uniquely captures key enterprise characteristics including data source fragmentation, access control hierarchies, and cross-functional workflows.Additionally, we provide a novel data generation pipeline that creates internally consistent enterprise tasks from organizational metadata.Experiments with state-of-the-art LLM agents demonstrate that even the most capable models achieve only 41.8% task completion, highlighting significant opportunities for improvement in enterprise-focused AI systems.

Ankush Agarwal, Chaitanya Devaguptapu, Ganesh S. Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers). 2025.

Given unstructured text, Large Language Models (LLMs) are adept at answering simple (single-hop) questions.However, as the complexity of the questions increase, the performance of LLMs degrade.We believe this is due to the overhead associated with understanding the complex question followed by filtering and aggregating unstructured information in the raw text.Recent methods try to reduce this burden by integrating structured knowledge triples into the raw text, aiming to provide a structured overview that simplifies information processing.However, this simplistic approach is query-agnostic and the extracted facts are ambiguous as they lack context.To address these drawbacks and to enable LLMs to answer complex (multi-hop) questions with ease, we propose to use a knowledge graph (KG) that is context-aware and is distilled to contain query-relevant information.The use of our compressed distilled KG as input to the LLM results in our method utilizing up to 67% fewer tokens to represent the query relevant information present in the supporting documents, compared to the state-of-the-art (SoTA) method.Our experiments show consistent improvements over the SoTA across several metrics (EM, F1, BERTScore, and Human Eval) on two popular benchmark datasets (HotpotQA and MuSiQue).

In the commercial aviation domain, there are a large number of documents, like accident reports of NTSB and ASRS, and regulatory directives ADs. There is a need for a system to efficiently access these diverse repositories to serve the demands of the aviation industry, such as maintenance, compliance, and safety. In this paper, we propose a Knowledge Graph (KG) guided Deep Learning (DL) based Question Answering (QA) system to cater to these requirements. We construct a KG from aircraft accident reports and contribute this resource to the community of researchers. The efficacy of this resource is tested and proved by the proposed QA system. Questions in Natural Language are converted into SPARQL (the interface language of the RDF graph database) queries and are answered from the KG. On the DL side, we examine two different QA models, BERT-QA and GPT3-QA, covering the two paradigms of answer formulation in QA. We evaluate our system on a set of handcrafted queries curated from the accident reports. Our hybrid KG + DL QA system, KGQA + BERT-QA, achieves 7% and 40.3% increase in accuracy over KGQA and BERT-QA systems respectively. Similarly, the other combined system, KGQA + GPT3-QA, achieves 29.3% and 9.3% increase in accuracy over KGQA and GPT3-QA systems respectively. Thus, we infer that the combination of KG and DL is better than either KG or DL individually for QA, at least in our chosen domain.

Joggled fishplate is an essential component that helps with emergency repairs to fractured sections of the rail. The monitoring of this component is essential in order to maintain steadiness of the train. Manual operations involve railroad personnel. These operations are subjective and time-consuming. Consequently, replacement of manual inspections with unmanned vehicles such as drones is well suited for fast and efficient maintenance operations. Therefore, in this work, we have proposed a critical analysis for development of a novel feature-based template matching technique for efficient and fast joggled fishplate detection in drone data. The appropriate feature is chosen for performing sliding window-based template matching using Euclidean distance. Non-maximum suppression has been used for reduction of false detections. The specificity and sensitivity values are calculated as 0.98 and 0.79 respectively.

Satellite imagery has been used in most of the applications to estimate, monitor and predict various situations and hurdles in the current scenario. According to the application, the satellite data is chosen on the basis of its spatial, spectral and temporal characteristics. From agriculture to the road highways construction and their health monitoring, it is playing a major role by providing its features details from a high level to medium and low level. In the context of agriculture, there are many ways to find crop health like visual interpretation, health parameter, existing vegetation indices, etc. For the purpose of monitoring and predicting, we require band values from the satellite data which may lack in providing accurate values because of the several parameters that affect it by which there is a possibility of error because they have to be calibrated. Therefore in this paper, a machine learning based calibration approach is proposed by considering multispectral drone data as in-situ data. The methodology is developed and tested for Sentinel-2 data. It is observed that the proposed calibration technique has good potential to calibrate satellite data.