Pop songs will get shorter on average by the end of this decade because of faltering attention spans and'skipping culture' on streaming services, experts say. Attention spans of music fans has dropped from 12 seconds to eight since the year 2000, according to research from Samsung. As a result it's more important than ever for musicians to draw listeners in early, keep the overall length of a track short and'load choruses up front'. On music streaming services like Spotify, artists don't get royalties from a song being played if the listener doesn't get beyond the first 30 seconds. By 2030, it will therefore be more important than ever for songs to quickly progress to the next track on an album before listeners get bored, the experts say.

While the circumstances were hardly ideal, 2020 was a big year for cord cutting. Streaming services like Netflix and Disney saw their subscriber numbers soar as people looked for ways to pass the time at home, and the temporary suspension of live sports accelerated the decline of traditional pay TV bundles. While not coronavirus-related, this year introduced two major new streaming services in HBO Max and NBC's Peacock, and we saw some bold attempts to rethink the streaming device with Google's new Chromecast and the TiVo Stream 4K. I've been writing this weekly column (and newsletter) on cord-cutting through it all, so in accordance with annual tradition, I'd like to cap off 2020 by recounting my favorite developments of the year. Here are TechHive's fifth-annual cord-cutter awards: At the start of the year, I wrote that unified streaming TV guides would be one of cord-cutting's biggest trends, and no streaming device delivers on that idea quite like the Chromecast with Google TV.

In 2018, Sam Cole, a reporter at Motherboard, discovered a new and disturbing corner of the internet. A Reddit user by the name of "deepfakes" was posting nonconsensual fake porn videos using an AI algorithm to swap celebrities' faces into real porn. Cole sounded the alarm on the phenomenon, right as the technology was about to explode. A year later, deepfake porn had spread far beyond Reddit, with easily accessible apps that could "strip" clothes off any woman photographed. Since then deepfakes have had a bad rap, and rightly so.

This time of year, many people post their Spotify'Wrapped' lists, revealing which songs they played the most over the past 12 months. The online culture magazine The Pudding fed millions of indicators of what it deemed objectively good music, including Pitchfork reviews, record store recommendations'and subreddits you've never heard of.' In addition to rating your taste in tracks, the'How Bad Is Your Spotify?' app will tell you which songs you play too much and which artists you are obsessed with'to an uncomfortable extent.' Launching on Wednesday, it's proven successful enough to trend on Twitter. Logging into Spotify via the'How Bad is Your Spotify' allows the site to review and judge your'awful' taste in music Depending on how high traffic is, it may take a few moments for it to assess your taste in music.

The best book ever written about IT work or the computer industry will be 40 years old in August. Tracy Kidder's The Soul of a New Machine describes the work of Data General engineers to prototype a minicomputer, codenamed "Eagle," intended to halt the advance of the Digital Equipment Corporation's hugely successful VAX range. It won both the Pulitzer Prize and National Book Award for non-fiction, perhaps the two highest honors available for book-length journalism. Year after year, the book continues to sell and win new fans. Developers born since it was published often credit it with shaping their career choices or helping them appreciate the universal aspects of their own experiences. Soul's appeal has endured, even though what started out as a dispatch from a fast-growing firm building a piece of the future now reads as a time capsule from a lost world. Back in 1991 I read the book for an undergraduate class, typing my paper on a PC that was already more capable than Eagle yet cost 100 times less. So why are so many people still excited to relive the creation of a pitifully obsolete computer, designed by a team of obscure engineers for a long-forgotten company that never mattered very much anyway? Having spent almost 30 years now trying to take the book apart and figure out how it works, I think I have some answers. Paradoxically, the obscurity of Data General helps to explain the book's enduring power.

Yahoo Japan Corp. said Wednesday it will delete hateful and defamatory comments from all of its online posting sites with the help of artificial intelligence, beefing up efforts to tackle cyberbullying after the suspected suicide of a reality television show star. The operator of Yahoo online services is strengthening its monitoring of hateful posts following the death earlier this year of Hana Kimura, a cast member of the popular reality show "Terrace House," who had been the target of bullying on social media. Yahoo Japan will release a list of expressions that could be taken as malicious and clarify the criteria for deleting. Posts that are judged harmful by AI will be automatically removed. If a user known to have made inappropriate posts gets a different ID, Yahoo Japan will suspend them.

The prevalence of state-sponsored propaganda on the Internet has become a cause for concern in the recent years. While much effort has been made to identify state-sponsored Internet propaganda, the problem remains far from being solved because the ambiguous definition of propaganda leads to unreliable data labelling, and the huge amount of potential predictive features causes the models to be inexplicable. This paper is the first attempt to build a balanced dataset for this task. The dataset is comprised of propaganda by three different organizations across two time periods. A multi-model framework for detecting propaganda messages solely based on the visual and textual content is proposed which achieves a promising performance on detecting propaganda by the three organizations both for the same time period (training and testing on data from the same time period) (F1=0.869) and for different time periods (training on past, testing on future) (F1=0.697). To reduce the influence of false positive predictions, we change the threshold to test the relationship between the false positive and true positive rates and provide explanations for the predictions made by our models with visualization tools to enhance the interpretability of our framework. Our new dataset and general framework provide a strong benchmark for the task of identifying state-sponsored Internet propaganda and point out a potential path for future work on this task.

One of the grand challenges discussed during the Dagstuhl Seminar "Knowledge Graphs: New Directions for Knowledge Representation on the Semantic Web" and described in its report is that of a: "Public FAIR Knowledge Graph of Everything: We increasingly see the creation of knowledge graphs that capture information about the entirety of a class of entities. [...] This grand challenge extends this further by asking if we can create a knowledge graph of "everything" ranging from common sense concepts to location based entities. This knowledge graph should be "open to the public" in a FAIR manner democratizing this mass amount of knowledge." Although linked open data (LOD) is one knowledge graph, it is the closest realisation (and probably the only one) to a public FAIR Knowledge Graph (KG) of everything. Surely, LOD provides a unique testbed for experimenting and evaluating research hypotheses on open and FAIR KG. One of the most neglected FAIR issues about KGs is their ongoing evolution and long term preservation. We want to investigate this problem, that is to understand what preserving and supporting the evolution of KGs means and how these problems can be addressed. Clearly, the problem can be approached from different perspectives and may require the development of different approaches, including new theories, ontologies, metrics, strategies, procedures, etc. This document reports a collaborative effort performed by 9 teams of students, each guided by a senior researcher as their mentor, attending the International Semantic Web Research School (ISWS 2019). Each team provides a different perspective to the problem of knowledge graph evolution substantiated by a set of research questions as the main subject of their investigation. In addition, they provide their working definition for KG preservation and evolution.

The first rumblings came when Diana Prince's metallic boots crossed No Man's Land in 2017's Wonder Woman. The movement gathered steam when Carol Danvers fell through the roof of that Blockbuster two years later. Female-led superhero movies were finally here--and they were about to be huge. Among the many other trends that never really caught on in 2020, the Year of the Female Superhero never quite came to pass. On a different timeline, this year would have begun with Birds of Prey and ended with Angelina Jolie and team of immortal heroes saving the world in Eternals.

Neural networks with deep architectures have demonstrated significant performance improvements in computer vision, speech recognition, and natural language processing. The challenges in information retrieval (IR), however, are different from these other application areas. A common form of IR involves ranking of documents -- or short passages -- in response to keyword-based queries. Effective IR systems must deal with query-document vocabulary mismatch problem, by modeling relationships between different query and document terms and how they indicate relevance. Models should also consider lexical matches when the query contains rare terms -- such as a person's name or a product model number -- not seen during training, and to avoid retrieving semantically related but irrelevant results. In many real-life IR tasks, the retrieval involves extremely large collections -- such as the document index of a commercial Web search engine -- containing billions of documents. Efficient IR methods should take advantage of specialized IR data structures, such as inverted index, to efficiently retrieve from large collections. Given an information need, the IR system also mediates how much exposure an information artifact receives by deciding whether it should be displayed, and where it should be positioned, among other results. Exposure-aware IR systems may optimize for additional objectives, besides relevance, such as parity of exposure for retrieved items and content publishers. In this thesis, we present novel neural architectures and methods motivated by the specific needs and challenges of IR tasks.