Information extraction (IE) from Visually Rich Documents (VRDs) containing layout features along with text is a critical and well-studied task. Specialized non-LLM NLP-based solutions typically involve training models using both textual and geometric information to label sequences/tokens as named entities or answers to specific questions. However, these approaches lack reasoning, are not able to infer values not explicitly present in documents, and do not generalize well to new formats. Generative LLMs-based approaches proposed recently are capable of reasoning, but struggle to comprehend clues from document layout especially in previously unseen document formats, and do not show competitive performance in heterogeneous VRD benchmark datasets. In this paper, we propose BLOCKIE, a novel LLM-based approach that organizes VRDs into localized, reusable semantic textual segments called \textit{semantic blocks}, which are processed independently. Through focused and more generalizable reasoning,our approach outperforms the state-of-the-art on public VRD benchmarks by 1-3% in F1 scores, is resilient to document formats previously not encountered and shows abilities to correctly extract information not explicitly present in documents.

Text-to-SQL bridges the gap between natural language and structured database language, thus allowing nontechnical users to easily query databases. Traditional approaches model text-to-SQL as a direct translation task, where a given Natural Language Query (NLQ) is mapped to an SQL command. Recent advances in large language models (LLMs) have significantly improved translation accuracy, however, these methods all require that the target database is pre-specified. This becomes problematic in scenarios with multiple extensive databases, where identifying the correct database becomes a crucial yet overlooked step. In this paper, we propose a three-stage end-to-end text-to-SQL framework to identify the user’s intended database before generating SQL queries. Our approach leverages LLMs and prompt engineering to extract implicit information from natural language queries (NLQs) in the form of a ruleset. We then train a large db_id prediction model, which includes a RoBERTa-based finetuned encoder, to predict the correct Database identifier (db_id) based on both the NLQ and the LLM-generated rules. Finally, we refine the generated SQL by using critic agents to correct errors. Experimental results demonstrate that our framework outperforms the current state-of-the-art models in both database intent prediction and SQL generation accuracy.

Invoices and receipts submitted by employees are visually rich documents (VRDs) with textual, visual and layout information. To protect against the risk of fraud and abuse, it is crucial for organizations to efficiently extract desired information from submitted receipts. This helps in the assessment of key factors such as appropriateness of the expense claim, adherence to spending and transaction policies, the validity of the receipt, as well as down-stream anomaly detection at various levels. These documents are heterogeneous, with multiple formats and languages, uploaded with different image qualities, and often do not contain ground truth labels for the efficient training of models. In this paper we propose Task Aware Instruction-based Labelling (TAIL), a method for synthetic label generation in VRD corpuses without labels, and fine-tune a multimodal Visually Rich Document Understanding Model (VRDU) on TAIL labels using response-based knowledge distillation without using the teacher model's weights or training dataset to conditionally generate annotations in the appropriate format. Using a benchmark external dataset where ground truth labels are available, we demonstrate conditions under which our approach performs at par with Claude 3 Sonnet through empirical studies. We then show that the resulting model performs at par or better on the internal expense documents of a large multinational organization than state-of-the-art LMM (large multi-modal model) Claude 3 Sonnet while being 85% less costly and ~5X faster, and outperforms layout-aware baselines by more than 10% in Average Normalized Levenshtein Similarity (ANLS) scores due to its ability to reason and extract information from rare formats. Finally, we illustrate the usage of our approach in overpayment prevention.

Summarization of large corpus of text is an important and hard problem. The amount of text information is increasing at an exponential rate which makes it consumption and organization difficult. This data also involves information obtained from transcription, chat along with traditional long form content such as books, periodicals, web pages etc. In this paper, we focus on long conversations obtained via transcription of multi speaker audio calls. Most of the traditional works text summarization focus on structured text summarization, however, we focus on multi turn conversation summarization. Many recent works perform model finetuning which improves the abstractive summarization by multi folds, it can still be improved further by incor-porating text into the encoder framework. Therefore, in this work we have studied the discourse-based saliency model that identifies the important segments within a text conversation and combines them adaptively to generate summary of the conversations. We demonstrate that the proposed method achieves state of the art re-sults on public benchmarks.

In this paper we solve the problem of detecting relationships between pairs of objects in an image. We develop spatially aware word embeddings using scene graphs and use joint feature representations containing visual, spatial and semantic embeddings from the input images to train a deep network on the task of relationship detection. Further, we propose to utilize context aligned scene graph embeddings from the train set, without requiring explicit availability of scene graphs at test time. We show that the proposed method outperforms the state-of-the-art methods for predicate detection and provides competing results on relationship detection. We also show the generalization ability of the proposed method by performing predictions under zero shot settings. Further, we also provide an exhaustive empirical evaluation on each component of the proposed network.

Challenges presented by ever increasing volumes of smart objects are huge in terms of security and privacy of data generated out of these objects. These challenges range from the usage scenario with other objects and dedicated servers which act as a middleman among various networks such as internet and wireless sensor network. In this paper, attempt has been made to highlight a simple use case of a composite server in a XEN-based virtualised system. Security threats like sniffing and spoofing are isolated and analysed with respect to a novel key transfer method between host and guest operating system in XEN-based virtualised system. In addition, secure and trust model to cater the said security threats like sniffing and spoofing is also presented. The performance of the proposed system in terms of CPU usage and network bandwidth is also shown.

Increasingly sophisticated image editing tools make it easy to modify images. Often these modifications are elaborate, convincing, and undetectable by even careful human inspection. These considerations have prompted the development of forensic algorithms and approaches to detect modifications done to an image. However, these detectors are model-driven (i.e., manipulation-specific) and the choice of a potent detector requires knowledge of the type of manipulation, something that cannot be known ( a priori ). Thus, the latest effort is directed towards developing model-free (i.e., generalized) detectors capable of detecting multiple manipulation types. In this article, we propose a novel detector capable of exposing seven different manipulation types in low-resolution compressed images. Our proposed approach is based on a two-layer convolutional neural network (CNN) to extract frequency domain features of image median filtered residual that are classified using two different classifiers—softmax and extremely randomized trees. Extensive experiments demonstrate the efficacy of proposed detector over existing state-of-the-art detectors.

In this paper, we propose to synthesize natural images from a set of input objects. The proposed technique generates a scene which has high correlation with the provided set of input objects while also maintaining the natural placement of objects within the scene. The technique constitutes of a generative adversarial network trained on a large corpus of objects and natural scenes. This is in contrast with earlier works where the objective was to generate a natural scene from a noise vector or conditioning the network over a variable. However, such methods have limitations in their ability to control the objects within the generated images. On the contrary, we show that by training a Generative Adversarial Network with raw image pixels as input, we can generate scenes which constitute the objects as well as generate the surrounding environment suitable for the combination of the input objects. We provide qualitative and quantitative results on challenging MS-COCO dataset to show the effectiveness of the proposed technique.

Image annotation (tag) and classification play a critical role in many computer vision applications, such as image retrieval, scene understanding, scene description etc. While, databases such as ImageNet have high quality labels for images, in real world, a large number of images have missing labels or tags that completely describe the contents of an image. To solve this problem, in this paper, we work on the hypothesis that class and tag information are correlated and propose a joint optimization for image classification and annotation. We construct a unified cost function to learn the class scoring vectors as well as tag scoring vectors. The proposed approach achieves state-of-the-art results on benchmark datasets for joint tag prediction and classification.