In attempts to develop sample-efficient algorithms, researcher have explored myriad mechanisms for collecting and exploiting feature feedback, auxiliary annotations provided for training (but not test) instances that highlight salient evidence. Examples include bounding boxes around objects and salient spans in text. Despite its intuitive appeal, feature feedback has not delivered significant gains in practical problems as assessed on iid holdout sets. However, recent works on counterfactually augmented data suggest an alternative benefit of supplemental annotations: lessening sensitivity to spurious patterns and consequently delivering gains in out-of-domain evaluations. Inspired by these findings, we hypothesize that while the numerous existing methods for incorporating feature feedback have delivered negligible in-sample gains, they may nevertheless generalize better out-of-domain. In experiments addressing sentiment analysis, we show that feature feedback methods perform significantly better on various natural out-of-domain datasets even absent differences on in-domain evaluation. By contrast, on natural language inference tasks, performance remains comparable. Finally, we compare those tasks where feature feedback does (and does not) help.

The residents of a relatively quiet neighborhood in San Francisco are seeing an influx of self-driving visitors. A steady stream of automated Waymo cars has reportedly been flowing into a dead end in the city's Richmond District, disrupting routes for passengers and forcing them to turn the car around themselves, local news outlet KPIX reported. And based on what neighbors are saying, it's not just one or two vehicles coming by. "There are some days where it can be up to 50," nearby resident Jennifer King told KPIX. And we're all working from home, so this is what we hear."

I believe anything that can replace the experience of reading can replace the profession of writing for reading. Audios and Videos have the most potential in achieving that. Continuity of the radio school of Information, the audiobooks, while may not be even close to print-books, have already fared better than ebooks. Podcasts are also doing good and there is an ever improvement on AI reading. There are innovations taking place in audiobooks and the fact that younger people tend to prefer it more, proves that audiobooks alongside cousin audio formats and styles are serious contenders to the throne of read-writing-hood. Along with audial formats, videos are also doing pretty good in providing information typically reserved for reading.

By now, most of us are familiar with the concept of AI. As one of the most popular technologies making rounds today, AI is talked about pretty often in mainstream media. While most of us have already heard of a few interesting applications of AI in different fields, many aren't aware that it has practical use in the world of entertainment! Entertainment and technology are closely intertwined, so when one changes, the other usually follows. With that in mind, here's how AI is used to improve some everyday entertainment channels we all know and love.

Bowers & Wilkins has launched a new version of its iconic Zeppelin speaker, and the company says it was re-imagined for the streaming age. The audio device manufacturer describes the new Zeppelin as "smarter and more flexible" than its predecessors, with built-in support for Amazon's Alexa voice assistant, so users can simply ask it to play whatever they want instead of using its physical buttons. In addition, B&W plans to give it multi-room capability in early 2022 through a software update. Once that arrives, users will be able to link several Zeppelins together or link a Zeppelin with other B&W speakers in a multi-room environment. For now, the new model supports AirPlay 2 and aptX Adaptive Bluetooth to give both iOS and Android users an easy way to stream audio from their devices.

We introduce Ego4D, a massive-scale egocentric video dataset and benchmark suite. It offers 3,025 hours of daily-life activity video spanning hundreds of scenarios (household, outdoor, workplace, leisure, etc.) captured by 855 unique camera wearers from 74 worldwide locations and 9 different countries. The approach to collection is designed to uphold rigorous privacy and ethics standards with consenting participants and robust de-identification procedures where relevant. Ego4D dramatically expands the volume of diverse egocentric video footage publicly available to the research community. Portions of the video are accompanied by audio, 3D meshes of the environment, eye gaze, stereo, and/or synchronized videos from multiple egocentric cameras at the same event. Furthermore, we present a host of new benchmark challenges centered around understanding the first-person visual experience in the past (querying an episodic memory), present (analyzing hand-object manipulation, audio-visual conversation, and social interactions), and future (forecasting activities). By publicly sharing this massive annotated dataset and benchmark suite, we aim to push the frontier of first-person perception. Project page: https://ego4d-data.org/

In many contexts, lying -- the use of verbal falsehoods to deceive -- is harmful. While lying has traditionally been a human affair, AI systems that make sophisticated verbal statements are becoming increasingly prevalent. This raises the question of how we should limit the harm caused by AI "lies" (i.e. falsehoods that are actively selected for). Human truthfulness is governed by social norms and by laws (against defamation, perjury, and fraud). Differences between AI and humans present an opportunity to have more precise standards of truthfulness for AI, and to have these standards rise over time. This could provide significant benefits to public epistemics and the economy, and mitigate risks of worst-case AI futures. Establishing norms or laws of AI truthfulness will require significant work to: (1) identify clear truthfulness standards; (2) create institutions that can judge adherence to those standards; and (3) develop AI systems that are robustly truthful. Our initial proposals for these areas include: (1) a standard of avoiding "negligent falsehoods" (a generalisation of lies that is easier to assess); (2) institutions to evaluate AI systems before and after real-world deployment; and (3) explicitly training AI systems to be truthful via curated datasets and human interaction. A concerning possibility is that evaluation mechanisms for eventual truthfulness standards could be captured by political interests, leading to harmful censorship and propaganda. Avoiding this might take careful attention. And since the scale of AI speech acts might grow dramatically over the coming decades, early truthfulness standards might be particularly important because of the precedents they set.

Sun Glare widely exists in the images captured by unmanned ground and aerial vehicles performing in outdoor environments. The existence of such artifacts in images will result in wrong feature extraction and failure of autonomous systems. Humans will try to adapt their view once they observe a glare (especially when driving), and this behavior is an essential requirement for the next generation of autonomous vehicles. The source of glare is not limited to the sun, and glare can be seen in the images captured during the nighttime and in indoor environments, which is due to the presence of different light sources; reflective surfaces also influence the generation of such artifacts. The glare's visual characteristics are different on images captured by various cameras and depend on several factors such as the camera's shutter speed and exposure level. Hence, it is challenging to introduce a general - robust and accurate - algorithm for glare detection that can perform well in various captured images. This research aims to introduce the first dataset for glare detection, which includes images captured by different cameras. Besides, the effect of multiple image representations and their combination in glare detection is examined using the proposed deep network architecture. The released dataset is available at https://github.com/maesfahani/glaredetection

While the world impatiently awaited Kanye West's new album, "DONDA", to drop, Wesam Jawich, a software engineer at Google, had an idea. What if we could just ask the outspoken artist when the album was dropping? So the idea was born to create an AI that would simulate a text conversation with Ye. The first step was to create a dataset of Kanye West dialogue to train GPT-J on. The dataset used was a compilation of Kanye interview transcripts, tweets, lyrics, and manufactured conversations.

On Wednesday, October 13 at 8:30am CT, William Shatner, who starred as Captain Kirk in Star Trek: The Original Series, will be going where no Hollywood star has gone before -- a suborbital sojourn that will take him 66 miles to the edge of space where he will be able to marvel at the curvature of earth and enjoy zero gravity weightlessness with his Blue Origin crew members. The entire experience is expected to last about 10 minutes and will be similar to the ride that Blue Origin and Amazon founder Jeff Bezos took this past summer with his brother and crew. Shatner might be the oldest person to rocket to space at 90 years old, but he's not the first actor to go. Feature filmmakers from Russia landed on the International Space Station last week beating Tom Cruise to bragging rights. The actor has been working on a $200 million Universal Studios film with SpaceX founder Elon Musk which NASA tweeted last year is expected to be shot on the space station.