1,720,994 research outputs found
ChatGPT Act as an Intelligence Officer
No full textThis paper explores the potential of Large Language Models (LLMs), specifically ChatGPT, to facilitate the work of intelligence officers. LLMs have demonstrated remarkable capabilities in natural language processing and generation, making them valuable tools for information retrieval, analysis, and decision-making. This paper presents an initial overview of the capabilities of LLMs in intelligence operations. Through a case study on ChatGPT, we illustrate how intelligence officers can leverage LLMs to enhance their work, including data collection, fusion, analysis, report generation, and information sharing. The results showcase the vast potential and the limits of LLMs in supporting intelligence officers in their complex and demanding tasks
Fusing contextual word embeddings for concreteness estimation
No full textNatural Language Processing (NLP) has a long history, and recent research has focused in particular on encoding meaning in a computable way. Word embeddings have been used for this specific purpose, allowing language tasks to be treated as mathematical problems. Real valued vectors have been generated or employed as word representations for several NLP tasks. In this work, different types of pre-trained word embeddings are fused together to estimate word concreteness. In the evaluation of this task, we have taken into account how much contextual information can affect final results, and also how to properly fuse different word embeddings in order to maximize their performance. The best architecture in our study surpasses the winning solution in the Evalita 2020 competition for the word concreteness task
Probing the Consistency of Situational Information Extraction with Large Language Models: A Case Study on Crisis Computing
No full textThe recently introduced foundation models for language modeling, also known as Large Language Models (LLMs), have demonstrated breakthrough capabilities on text summarization and contextualized natural language processing. However, these also suffer from inherent deficiencies like the occasional generation of factually wrong information, known as hallucinations, and a weak consistency of produced answers strongly varying with the exact phrasing of their input query, i.e., prompt. Hence, this raises the question whether and how LLMs could replace or complement traditional information extraction and fusion modules in information fusion pipelines involving textual input sources. We empirically examine this question on a case study from crisis computing, taken from the established CrisisFacts benchmark dataset, by probing an LLM’s situation understanding and summarization capabilities on the target task of extracting information relevant for establishing crisis situation awareness from social media corpora. Since social media messages are exchanged in real-time, typically targeting human readers aware of the situational context, this domain represents a prime testbed for evaluating LLMs’ situational information extraction capabilities. In this work, we specifically investigate the consistency of extracted information across different model configurations and different but semantically similar prompts, which represents a crucial prerequisite for a reliable and trustworthy information extraction component
Multimodal feature fusion for concreteness estimation
No full textIn recent years the idea of fusing diverse type of information has often been employed to solve various Deep Learning tasks. Whether these regard an NLP problem or a Machine Vision one, the concept of using more inputs of the same type has been the basis of many studies. Considering NLP problems, attempts of different word embeddings have already been tried, managing to make improvements to the most common benchmarks. Here we want to explore the combination not only of different types of input together, but also different data modalities. This is done by fusing two popular word embeddings together, mainly ELMo and BERT, with other inputs that embed a visual description of the analysed text. Doing so, different modalities -textual and visual- are both employed to solve a textual problem, a concreteness task. Multimodal feature fusion is here explored through several techniques: input redundancy, concatenation, average, dimensionality reduction and augmentation. By combining these techniques it is possible to generate different vector representations: the goal is to understand which feature fusion techniques allow to obtain more accurate embeddings
Survey on data fusion approaches for fall-detection
No full textHuman fall detection is a critical research area focused on developing methods and systems that can automatically detect and recognize falls, particularly among the elderly and individuals with disabilities. Falls are a major cause of injuries and deaths among these populations, and timely intervention can reduce the severity of consequences. This article presents a comprehensive review of fall detection systems, emphasizing the use of cutting-edge technologies such as deep learning, sensor fusion, and machine learning. The research explores a variety of methodologies and strategies employed in fall detection systems, including the integration of wearable sensors, smartphones, and cameras. By examining various fall detection techniques and their experimental results, the article highlights the effectiveness of these systems in identifying and classifying falls. The study also addresses the challenges and limitations associated with fall detection systems, emphasizing the need for ongoing research and advancements. In summary, this research contributes to the development of advanced fall detection systems, demonstrating their potential to improve the quality of life for the elderly, alleviate healthcare burdens, and provide reliable solutions for fall detection and classification
Ensemble of KalmanNets for Maneuvering Target Tracking
No full textTracking a maneuvering target requires the modeling of the target's movements by multiple pre-defined mathematical models. However, the uncertainty in the target's dynamics can lead traditional model-based (MB) tracking algorithms to significant performance degradation when model mismatch occurs. To tackle this problem, we propose the use of a Recurrent Neural Network (RNN) for the purpose of learning complex target dynamics. Following the recent advances in state estimation provided by KalmanNet, a neural network-aided Kalman Filter, the proposed approach aims to exploit its tracking performance in a multiple model schema to compensate for model mismatch across maneuvers, leading to a more prompt response to motion switches. The results over a simulated set of maneuvering target trajectories demonstrate the potential of the proposed approach over the MB solution
Tracking human motion from monocular sequences
No full textIn recent years, analysis of human motion has become an increasingly relevant research topic with applications as diverse as animation, virtual reality, security, and advanced human-machine interfaces. In particular, motion capture systems are well known nowadays since they are used in the movie industry. These systems require expensive multi-camera setups or markers to be worn by the user. This paper describes an attempt to provide a markerless low cost and real-time solution for home users. We propose a novel approach for robust detection and tracking of the user's body joints that exploits different algorithms as different sources of information and fuses their estimates with particle filters. This system may be employed for real-time animation of VRML or X3D avatars using an off-the-shelf digital camera and a standard PC
Cascaded online boosting
No full textIn this paper, we propose a cascaded version of the online boosting algorithm to speed-up the execution time and guarantee real-time performance even when employing a large number of classifiers. This is the case for target tracking purposes in computer vision applications. We thus revise the online boosting framework by building on-the-fly a cascade of classifiers dynamically for each new frame. The procedure takes into account both the error and the computational requirements of the available features and populates the levels of the cascade accordingly to optimize the detection rate while retaining real-time performance. We demonstrate the effectiveness of our approach on standard datasets
Towards Neural Situation Evolution Modeling: Learning a Distributed Representation for Predicting Complex Event Sequences
No full textIn real-world monitoring tasks, a situation can be understood as a sequence of causally related events of interest. In road traffic control, such a situation could be a rear-end collision at the end of a traffic jam, which worsens congestion and requires clearing operations and potentially rerouting. Whereas conventional event sequence prediction focuses on sequences of individual events 〈 e1, ⋯, en〉, evolving situations thus can be conceived as sequences of states composed of multiple concurrent events, i.e., complex events: 〈e1, ⋯, em, ⋯, el, ⋯, en〉. Situation (evolution) prediction thus requires learning a transition model for these complex events to provide the expectations for potential successor event types. In previous work, this was represented by a Markov Chain defined on the observed complex events. However, using the entire event composite as 'atomic' situation state representation does not allow capturing patterns between its individual events (e.g., events of type 'accident' share similar successor event types across different event composites), nor generalizing behaviors between similar event types or incorporating additional features. Hence, we propose a neural modeling approach to learn a distributed representation of a given situation dataset. By encoding the input states as conjunction of their individual comprised events, the devised model can learn associations (i.e., enable an 'information flow') between individual event types, allowing to capture similar behaviors across different situations. We test our approach on both synthetic and real-world datasets
- …
