459 research outputs found
SNOW 2014 Data Challenge
No full text<p>Resources and data from the SNOW 2014 Data Challenge. If you use these resources for your research, please cite:</p>
<p>Symeon Papadopoulos, David Corney, Luca Aiello (2014). SNOW 2014 Data Challenge: Assessing the Performance of News Topic Detection Methods in Social Media. <em>Proceedings of the SNOW 2014 Data Challenge</em>, April 2014</p>
<p>For more information see: http://www.socialsensor.eu/use-cases/evaluation/122-snow2014-data-challenge</p
Social computing for verifying social media content in breaking news
No full textSocial media is the place to go for both journalists and the general public when news events break, offering a real-time source of eyewitness images and videos through platforms like YouTube, Instagram, and Periscope. Yet, the value of such content as a means of documenting and disseminating breaking news is compromised by the increasing amount of content misuse and false claims in social media. To this end, cost-effective social computing solutions for user-generated content verification are crucial for retaining the value and trust in social media for breaking news
Location extraction from social media: geoparsing, location disambiguation and geotagging
No full textLocation extraction, also called toponym extraction, is a field covering geoparsing, extracting spatial representations from location mentions in text, and geotagging, assigning spatial coordinates to content items. This paper evaluates five ‘best of class’ location extraction algorithms. We develop a geoparsing algorithm using an OpenStreetMap database, and a geotagging algorithm using a language model constructed from social media tags and multiple gazetteers. Third party work evaluated includes a DBpedia-based entity recognition and disambiguation approach, a named entity recognition and Geonames gazetteer approach and a Google Geocoder API approach. We perform two quantitative benchmark evaluations, one geoparsing tweets and one geotagging Flickr posts, to compare all approaches. We also perform a qualitative evaluation recalling top N location mentions from tweets during major news events. The OpenStreetMap approach was best (F1 0.90+) for geoparsing English, and the language model approach was best (F1 0.66) for Turkish. The language model was best (F1@1km 0.49) for the geotagging evaluation. The map-database was best (R@20 0.60+) in the qualitative evaluation. We report on strengths, weaknesses and a detailed failure analysis for the approaches and suggest concrete areas for further research
Proceedings of the 1st Workshop on AI bias: Measurements, Mitigation, Explanation Strategies co-located with the AI Fairness Cluster Inaugural Conference 2024
No full textThere were 24 papers submitted for peer-review to this workshop. Out of these, 6 papers were accepted for this volume, 6 as regular papers and 0 as short papers
Integration of Semantic, Metadata and Image Search Engines with a Text Search Engine for Patent Retrieval
No full textThe combination of different search techniques can improve the
results given by each one. In the ongoing R&D project PATExpert1, four
different search techniques are combined to perform a patent search. These
techniques are: metadata search, keyword-based search, semantic search and
image search. In this paper we propose a general architecture based on web
services where each tool works in its own domain and provides a set of basic
functionalities to perform the retrieval. To be able to combine the results from
the four search engines, these must be fuzzy (using a membership function or
similarity grade). We focus on how the fuzzy results can be obtained from each
technique, and how they can then be combined. This combination must take
into account the query, the similarity of the patent to each part of the query, and
the confidence on the technique
Valuation of Netflix, Inc.
Get PDFΔιπλωματική εργασία--Πανεπιστήμιο Μακεδονίας, Θεσσαλονίκη, 2021.Rapid technological developments are occurring quicker than ever before in the twenty-first century. COVID-19 further exacerbated this issue by emphasizing the continuous changing of schemes in our society and economy. Enterprises, being one of the primary engines of the economy, must constantly change by efficiently adjusting to the new realities enabled by technological advancement. On the other hand, markets are the most accessible they have ever been in human history. Nowadays, anybody may access the markets using a smartphone app. Despite this, the sheer quantity of analysts and specialists throughout the world has made it exceedingly difficult for any individual investor to identify and pick firms that may provide considerable potential. This incident highlights the significance of recognizing a company's true value and the correct procedure for estimating it.
Each company's value reflects its ability to generate profits for its owners. The evaluation of an entity's worth is a complex process that is governed by each individual's requirements. Its usefulness becomes important during the investment decision-making process.
The purpose of this master's thesis is to calculate a theoretical equity value for Netflix, Inc and, as a result, its value per share. By discounting projected future cash flows to present value, the enterprise value was calculated.
The dissertation presents and analyzes Netflix and the entertainment business from a strategic standpoint. The strategic analytical insight enabled a financial statement study of Netflix and the sector, followed by a forecast of future performance. The approach utilized was the Discounted Cash Flow model.
Finally, the ‘fair' share value was compared to the actual stock market price to determine if its value is underpriced or overvalued and to effectively estimate its potential. The conclusion falls on a final value estimate of $720.18 per share. Thus, suggesting a buy recommendation
Training in Co-Creation as a Methodological Approach to Improve AI Fairness
Get PDFParticipatory design (PD) and co-creation (Co-C) approaches to building Artificial Intelligence (AI) systems have become increasingly popular exercises for ensuring greater social inclusion and fairness in technological transformation by accounting for the experiences of vulnerable or disadvantaged social groups; however, such design work is challenging in practice, partly because of the inaccessible domain of technical expertise inherent to AI design. This paper evaluates a methodological approach to make addressing AI bias more accessible by incorporating a training component on AI bias in a Co-C process with vulnerable and marginalized participant groups. This was applied by socio-technical researchers involved in creating an AI bias mitigation developer toolkit. This paper’s analysis emphasizes that critical reflection on how to use training in Co-C appropriately and how such training should be designed and implemented is necessary to ensure training allows for a genuinely more inclusive approach to AI systems design when those most at risk of being adversely affected by AI technologies are often not the intended end-users of said technologies. This is acutely relevant as Co-C exercises are increasingly used to demonstrate regulatory compliance and ethical practice by powerful institutions and actors developing AI systems, particularly in the ethical and regulatory environment coalescing around the European Union’s recent AI Act
Zero-Shot Content-Based Crossmodal Recommendation System
Get PDFInformation Recommendation (IR) systems are conventionally designed to operate within a single modality at a time, such as Text2Text or Image2Image. However, the concept of cross-modality aims to facilitate a versatile recommendation experience across different modalities, such as Text2Image. In recent years, significant strides have been made in developing neural recommender models that are multimodal and capable of generalizing across a broad spectrum of domains in a zero-shot manner, thanks to the robust representation capabilities of neural networks. These architectures enable the generation of embeddings for assets (i.e., content uploaded on a platform by users), presenting a concise representation of their semantics and allowing for comparisons through similarity ranking. In this paper, we present ZCCR, a Zero-shot Content-based Crossmodal Recommendation System that leverages knowledge from large-scale pretrained Vision-Language Models (VLMs) such as CLIP and ALBEF to redefine the recommendation task as a zero-shot retrieval task, eliminating the need for labeled data or prior knowledge about the recommended content. Furthermore, ZCCR performs crossmodal similarity search on an optimized index, such as FAISS, to improve the speed of recommendation. The goal is to recommend to the user assets that are similar to those they have previously uploaded, commonly referred to as their user profile. Within the user profile, we identify "areas of interest"-groups of assets associated with specific user interests, such as cooking, sports, or cars. To identify these areas of interest and construct the search query for the retrieval operation, ZCCR employs an innovative use of Agglomerative Clustering. This technique groups user past assets by similarity without requiring prior knowledge of the number of clusters. Once the areas of interest, or clusters, are identified, the centroid is utilized as the search seed to find similar assets. Experimental results demonstrate the efficiency of the selected components in terms of search time and retrieval performance on modified MSCOCO and FLICKR30k datasets tailored for the recommendation task. Furthermore, ZCCR outperforms both a baseline tagging system (BT) and a more advanced tag system which utilizes a Large Language Model (LLM) to extract embeddings from tags. The results show that, even compared with the latter, embeddings directly extracted from raw assets yield superior outcomes compared to relying on intermediate tags generated by other tools. The code implementation to reproduce all experiments and results shown in this paper is provided at the following link: ZCCR-experiments
- …
