Explaining the output of a complex system, such as a Recommender System (RS), is becoming of utmost importance for both users and companies. In this paper we explore the idea that personalized explanations can be learned as recommendation themselves. There are plenty of online services where users can upload some photos, in addition to rating items. We assume that users take these photos to reinforce or justify their opinions about the items. For this reason we try to predict what photo a user would take of an item, because that image is the argument that can best convince her of the qualities of the item. In this sense, an RS can explain its results and, therefore, increase its reliability. Furthermore, once we have a model to predict attractive images for users, we can estimate their distribution. The paper includes a formal framework that estimates the authorship probability for a given pair (user, photo). To illustrate the proposal, we use data gathered from TripAdvisor containing the reviews (with photos) of restaurants in six cities of different sizes. Keywords: Recommender Systems, Personalization, Explainability, Photo, Collaborative 1. Introduction Explainable Artificial Intelligence (XAI) is becoming an important area of interest since explainability is increasingly necessary to meet stakeholder demands. In particular, the General Data Protection Regulation (GDPR) [29] of the European Union demands transparency in systems that take decisions affecting people, making explanations more needed than ever. Additionally, explanations may help increase the trust of users in AI algorithms, since people rely not only on their efficacy but also on the degree of understanding of the process they follow. Since they provide suggestions to users, explainability plays an important role on them.

Among the existing approaches for visual-based Recommender System (RS) explainability, utilizing user-uploaded item images as efficient, trustable explanations is a promising option. However, current models following this paradigm assume that, for any user, all images uploaded by other users can be considered negative training examples (i.e. bad explanatory images), an inadvertedly naive labelling assumption that contradicts the rationale of the approach. This work proposes a new explainer training pipeline by leveraging Positive-Unlabelled (PU) Learning techniques to train image-based explainer with refined subsets of reliable negative examples for each user selected through a novel user-personalized, two-step, similarity-based PU Learning algorithm. Computational experiments show this PU-based approach outperforms the state-of-the-art non-PU method in six popular real-world datasets, proving that an improvement of visual-based RS explainability can be achieved by maximizing training data quality rather than increasing model complexity.

AI expert Marva Bailer explains how, even though there are currently laws in place, the average person has more access than ever to create deepfakes of celebrities. "Elvis" director Baz Luhrmann is not afraid of artificial intelligence so much as he worries about the lack of regulation over the technology. In an interview with Sky News, Luhrmann admitted he was not "personally frightened of AI, but having worked with a very, very smart robot named Ai-Da, and having formed a relationship with her, she would tell you, and I would agree, we are way behind in terms of governance of AI." Earlier this year, Luhrmann partnered with Bombay Sapphire on its "Saw This Made This" campaign, which used an AI robot artist, named Ai-Da, to create art pieces live at exhibitions in London and New York inspired by submissions from human creators. Luhrmann also praised the writers and actors strikes that took place over the summer and fall, with the use of AI being a major issue in negotiations. WHAT IS ARTIFICIAL INTELLIGENCE (AI)?

It's impossible to avoid Elvis Presley in Las Vegas: his image appears on street art and photographs, while impersonators can be found all over the strip. But starting on Friday, the King of Rock and Roll will get possibly his largest tribute yet: a video collage that renders him hundreds of times, projected hundreds of feet into the air, in incarnations young and old, gyrating and reclining, in bas relief and gold, all thanks to a technology created long after his death: generative AI. The video collage is the creation of the artist Marco Brambilla, the director of Demolition Man and Kanye West's "Power" music video, among many other art projects. Brambilla fed hours of footage from Presley's movies and performances into the AI model Stable Diffusion to create an easily searchable library to pull from, and then created surreal new images by prompting the AI model Midjourney with questions like: "What would Elvis look like if he were sculpted by the artist who made the Statue of Liberty?" The kaleidoscopic result, called "King Size," will make its debut as part of U2's concert performance at the opening night of the Sphere, a $2.3 billion entertainment venue that sits a block from the Las Vegas Strip and hopes to be the city's latest colossal entertainment mecca.

Deep learning has shown great potential in assisting radiologists in reading chest X-ray (CXR) images, but its need for expensive annotations for improving performance prevents widespread clinical application. Visual language pre-training (VLP) can alleviate the burden and cost of annotation by leveraging routinely generated reports for radiographs, which exist in large quantities as well as in paired form (image-text pairs). Additionally, extensions to localization-aware VLPs are being proposed to address the needs for accurate localization of abnormalities for computer-aided diagnosis (CAD) in CXR. However, we find that the formulation proposed by locality-aware VLP literature actually leads to a loss in spatial relationships required for downstream localization tasks. Therefore, we propose Empowering Locality of VLP with Intra-modal Similarity, ELVIS, a VLP aware of intra-modal locality, to better preserve the locality within radiographs or reports, which enhances the ability to comprehend location references in text reports. Our locality-aware VLP method significantly outperforms state-of-the art baselines in multiple segmentation tasks and the MS-CXR phrase grounding task. Qualitatively, we show that ELVIS focuses well on regions of interest described in the report text compared to prior approaches, allowing for enhanced interpretability.

The Monitor is a weekly column devoted to everything happening in the WIRED world of culture, from movies to memes, TV to Twitter. The Australian writer, director, and producer is known for his flashy, hyper-realistic style, and on this particular New York night he's in a sparse, brightly lit former taxi warehouse in Chelsea, talking to a robot. The bot's name is Ai-Da; she's a painter powered by artificial intelligence. Before Luhrmann took the stage next to her, she was doing a watercolor while people gawked and took photos. "Did you see Elvis, Ai-Da?" he asked.

On the floor of CES, LG's CLOi service robots got a lot of attention. But just across the parking lot from the Las Vegas Convention Center, two service robots--both Relay robots from San Jose-based Savioke--are quietly at work. These robots, tagged Elvis and Priscilla, are full-time employees of the Renaissance Hotel, and they aren't getting a lot of attention. When Priscilla navigated through the crowded lobby to make a delivery on Wednesday, only a few people pulled out cameras. Others casually brushed by, sometimes giving it a little pat as they passed.

This paper describes a novel method by which a spoken dialogue system can learn to choose an optimal dialogue strategy from its experience interacting with human users. The method is based on a combination of reinforcement learning and performance modeling of spoken dialogue systems. The reinforcement learning component applies Q-learning (Watkins, 1989), while the performance modeling component applies the PARADISE evaluation framework (Walker et al., 1997) to learn the performance function (reward) used in reinforcement learning. We illustrate the method with a spoken dialogue system named ELVIS (EmaiL Voice Interactive System), that supports access to email over the phone. We conduct a set of experiments for training an optimal dialogue strategy on a corpus of 219 dialogues in which human users interact with ELVIS over the phone. We then test that strategy on a corpus of 18 dialogues. We show that ELVIS can learn to optimize its strategy selection for agent initiative, for reading messages, and for summarizing email folders.