In this paper, we present ENTER, an interpretable Video Question Answering (VideoQA) system based on event graphs. Event graphs convert videos into graphical representations, where video events form the nodes and event-event relationships (temporal/causal/hierarchical) form the edges. This structured representation offers many benefits: 1) Interpretable VideoQA via generated code that parses event-graph; 2) Incorporation of contextual visual information in the reasoning process (code generation) via event graphs; 3) Robust VideoQA via Hierarchical Iterative Update of the event graphs. Existing interpretable VideoQA systems are often top-down, disregarding low-level visual information in the reasoning plan generation, and are brittle. While bottom-up approaches produce responses from visual data, they lack interpretability. Experimental results on NExT-QA, IntentQA, and EgoSchema demonstrate that not only does our method outperform existing top-down approaches while obtaining competitive performance against bottom-up approaches, but more importantly, offers superior interpretability and explainability in the reasoning process.

Continuous tracking of boxers across multiple training sessions helps quantify traits required for the well-known ten-point-must system. However, continuous tracking of multiple athletes across multiple training sessions remains a challenge, because it is difficult to precisely segment bout boundaries in a recorded video stream. Furthermore, re-identification of the same athlete over different period or even within the same bout remains a challenge. Difficulties are further compounded when a single fixed view video is captured in top-view. This work summarizes our progress in creating a system in an economically single fixed top-view camera. Specifically, we describe improved algorithm for bout transition detection and in-bout continuous player identification without erroneous ID updation or ID switching. From our custom collected data of ~11 hours (athlete count: 45, bouts: 189), our transition detection algorithm achieves 90% accuracy and continuous ID tracking achieves IDU=0, IDS=0.

ChatGPT is an amazing Chatbot with the ability, competency, and confidence to construct paragraphs, sentences, essays, and more. However, this optimism can be deceiving because it succumbs to several common misconceptions among the general public, even though experts know much better. Can an astrophysicist teach ChatGPT to acquire new information and absorb it such that it gives exact responses when it previously gave confident but mistaken ones? When we try to understand anything more intensely, we all end up in an awkward situation: we believe we know how something performs, only to explore that we are wrong. This entails not just learning what is true, but also why what we assumed was true was, in actuality, false, and how we can avoid making the same mistakes again.

The rogue-like elements are served up with a twist. Players can choose to either give up and start over, losing abilities that aren't permanently unlocked with experience points, or get back up and age up. Every decade lost brings increases to damage while decreasing available health. The challenge becomes finishing levels -- including grueling boss encounters -- without defeat. The game remembers the lowest age at which you beat a level when entering a new level, a sort of "save state" for future runs until you can beat levels flawlessly.

Free Guy, the new movie about a non-player character who discovers he's trapped inside a video game, is built around a series of fight scenes that are, for want of a better word, extremely video-gamey: Characters move stiffly, some have signature moves they repeat, and the laws of physics seem to have been imported from somewhere between The Matrix and Warner Bros. cartoons. We spoke to the film's fight coordinator, Freddy Bouciegues, to find out he choreographed human actors to fight like video game characters--and how that's different from his work on fights in actual video games. This conversation has been condensed and edited for clarity. I assume the two or three major fight sequences in the movie make up the bulk of your efforts, but were you also working on all random stunts going on in the background? Whenever Ryan Reynolds walks down the street, we get a taste of Grand Theft Auto-style mayhem caused by other players of the game-within-the-movie.

Language is contextual as meanings of words are dependent on their contexts. Contextuality is, concomitantly, a well-defined concept in quantum mechanics where it is considered a major resource for quantum computations. We investigate whether natural language exhibits any of the quantum mechanics' contextual features. We show that meaning combinations in ambiguous phrases can be modelled in the sheaf-theoretic framework for quantum contextuality, where they can become possibilistically contextual. Using the framework of Contextuality-by-Default (CbD), we explore the probabilistic variants of these and show that CbD-contextuality is also possible.

App builders commonly use security challenges, a form of step-up authentication, to add security to their apps. However, the ethical implications of this type of architecture has not been studied previously. In this paper, we present a large-scale measurement study of running an existing anti-fraud security challenge, Boxer, in real apps running on mobile devices. We find that although Boxer does work well overall, it is unable to scan effectively on devices that run its machine learning models at less than one frame per second (FPS), blocking users who use inexpensive devices. With the insights from our study, we design Daredevil, anew anti-fraud system for scanning payment cards that work swell across the broad range of performance characteristics and hardware configurations found on modern mobile devices. Daredevil reduces the number of devices that run at less than one FPS by an order of magnitude compared to Boxer, providing a more equitable system for fighting fraud. In total, we collect data from 5,085,444 real devices spread across 496 real apps running production software and interacting with real users.

Here's the thing: All the top-performing Medicare Advantage plans are using, today, right now, some form of advanced analytics and artificial intelligence (AI) to risk-stratify their populations and predict which members will, without intervention, become high cost in the near term. The idea is then to intervene to mitigate risk and stop bad things from happening--bad things that stink if you're the patient and also cost a lot if you're the plan. That's what population health management is all about, after all. Others using AI, right now, to do the kind of predictive analytics that you need to excel at pop health include PCP groups and other providers, mainly those at risk to manage populations or readmissions. In this health care podcast, I talk with Andrew Eye about AI.

Following consumer adoption and changing attitudes, purpose-built smart assistants like Alexa for Business are paving the way for AI at work. Smart assistants naturally complement intelligent smartphone apps, such as Edison, formerly EasilyDo. Using predictive analytics and deep learning, these smart apps extract meaningful, actionable data in real-time. For example, Edison proactively notifies you when it's time to leave based on traffic patterns and meeting start times. While these apps classify as consumer offerings, it's easy to understand why employees want these tools at work.

Many employers are using internships, boot camps and other after-school programs to net promising candidates before they earn their degrees, chief information officers say. "We regularly engage with students to help build brand awareness and candidate pipelines," said Ashley Pettit, CIO at State Farm Mutual Automobile Insurance Co. The Bloomington, Ill., insurer has about 6,000 technology workers, she said, and a recent online career fair it organized attracted more than 3,000 job seekers. Ms. Pettit was one of 30 information-technology executives who responded via email to CIO Journal's annual end-of-year questionnaire about hiring and other issues. CompTIA, an IT trade group, estimates that--despite ebbs and flows in the job market--the number of U.S. tech jobs is expected to grow 13.1% by 2026 from 2016, compared with 10.7% for all occupations.