Terms of the deal weren't disclosed. SparkCognition was founded in 2013 and makes machine-learning and AI technology.

In an expanded edition published in 1988, they responded to claims that their 1969 conclusions significantly reduced funding for neural network research: "Our version is that progress had already come to a virtual halt because of the lack of adequate basic theories… by the mid-1960s there had been a great many experiments with perceptrons, but no one had been able to explain why they were able to recognize certain kinds of patterns and not others."

Neural topic models have triggered a surge of interest in extracting topics from text automatically since they avoid the sophisticated derivations in conventional topic models. However, scarce neural topic models incorporate the word relatedness information captured in word embedding into the modeling process. To address this issue, we propose a novel topic modeling approach, called Variational Gaussian Topic Model (VaGTM). Based on the variational auto-encoder, the proposed VaGTM models each topic with a multivariate Gaussian in decoder to incorporate word relatedness. Furthermore, to address the limitation that pre-trained word embeddings of topic-associated words do not follow a multivariate Gaussian, Variational Gaussian Topic Model with Invertible neural Projections (VaGTM-IP) is extended from VaGTM. Three benchmark text corpora are used in experiments to verify the effectiveness of VaGTM and VaGTM-IP. The experimental results show that VaGTM and VaGTM-IP outperform several competitive baselines and obtain more coherent topics.

Interactive news recommendation has been launched and attracted much attention recently. In this scenario, user's behavior evolves from single click behavior to multiple behaviors including like, comment, share etc. However, most of the existing methods still use single click behavior as the unique criterion of judging user's preferences. Further, although heterogeneous graphs have been applied in different areas, a proper way to construct a heterogeneous graph for interactive news data with an appropriate learning mechanism on it is still desired. To address the above concerns, we propose a graph-based behavior-aware network, which simultaneously considers six different types of behaviors as well as user's demand on the news diversity. We have three main steps. First, we build an interaction behavior graph for multi-level and multi-category data. Second, we apply DeepWalk on the behavior graph to obtain entity semantics, then build a graph-based convolutional neural network called G-CNN to learn news representations, and an attention-based LSTM to learn behavior sequence representations. Third, we introduce core and coritivity features for the behavior graph, which measure the concentration degree of user's interests. These features affect the trade-off between accuracy and diversity of our personalized recommendation system. Taking these features into account, our system finally achieves recommending news to different users at their different levels of concentration degrees.

In high school, Mackenzie Leake shot a movie about being afraid to get her driver's license. "A very millennial subject," she jokes. It gave her a punishing lesson in editing video: Leake spent countless hours, over the course of weeks, "scrubbing" through her footage to find the best shots, then painstakingly assembling them. "It's a ton of grunt work," she notes. Now, seven years later, she's trying to accelerate the process.

According to a new Forrester report, "The Tech Executive's Primer On Data Science, Machine Learning, And AI, a lack of understanding is hampering the ability of business leaders to effectively deploy data science, machine learning and artificial intelligence projects to solve business problems. "All executives need to make strategic decisions about how and where to leverage these technologies, but few leaders have experience with them, so misconceptions abound, causing poor outcomes, wasted resources and resistance to future initiatives," the report said. The report defines data science as extracting meaning from data; machine learning as applying algorithms to data to train machine learning models; and artificial intelligence as an umbrella term for machine learning and automation methods used in new ways. To successfully deploy these technologies requires business and technology acumen and executive leadership, the report said. While technical experts can be hired, finding business executives who understand these complex, cutting-edge technologies is much harder. If it looks like you think AI "should" look, it's probably not. As smart as AI technologies like personal assistants or grammar-checkers appear, real-world AI does not exhibit anywhere near the intelligence and autonomy portrayed in the movies. "The actual advantages and disadvantages of ML and AI technologies vary so dramatically from popular perceptions that if an idea, proposed solution, or vendor offering looks like something a layperson would expect, it will be doomed to fail, is overly hyped or will have to rely on a person hiding behind a curtain," the report said. DSMLAI shouldn't start with just the end in mind or with what the AI and ML technologies can do. You have to meet in the middle. "Start purely with the business value and you'll choose use cases that play to AI's weaknesses and miss its strengths (think fully autonomous vehicles).

PornHub has delved into the past to remaster some of the oldest erotic movies in existence. The Remastured project (warning: that link will lead you to some very NSFW images) used AI to restore and colorize skin flicks from as far back as 125 years ago. The porn giant harnessed machine learning and 100,000 adult images and videos to teach the AI how to colorize the films (perhaps it also learned a thing or two about how people passed the time 100 years before smartphones). Several algorithms were used to restore the films with "limited human intervention," according to PornHub. The process started by reducing noise and sharpening and contrasting images.

I'll admit that I was more than a little skeptical when the Knight Foundation announced last week that it would award $3 million in grants to help local news organizations use artificial intelligence. My first reaction was that dousing the cash with gasoline and tossing a match would be just as effective. But then I started thinking about how AI has enhanced my own work as a journalist. For instance, just a few years ago I had two unappetizing choices after I recorded an interview: transcribing it myself or sending it out to an actual human being to do the work at considerable expense. Now I use an automated system, based on AI, that does a decent job at a fraction of the cost. Or consider Google, whose search engine makes use of AI.

New deepfake technology allows Robert De Niro to deliver his famous line from Taxi Driver in flawless German--with realistic lip movements and facial expressions. The AI software manipulates an actor's lips and facial expressions to make them convincingly match the speech of someone speaking the same lines in a different language. The artificial-intelligence-based tech could reshape the movie industry, in both alluring and troubling ways. The technology is related to deepfaking, which uses AI to paste one person's face onto someone else. It promises to allow directors to effectively reshoot movies in different languages, making foreign versions less jarring for audiences and more faithful to the original.

In February 2020, the digital media agency Space150 experimented with artificial intelligence machine learning programs to create the excellently titled "Jack Park Canny Dope Man," a banger of a hip-hop song in the vein of one of the genre's biggest names, Travis Scott. The neural networks used to craft the tune were trained on Scott's entire catalog, and the resulting beat and melody don't do anything to betray their artificial origins. The song lacks the awkward, clunking approximation of what a human operator would produce, something we so often see in language translation software, for example. No--it sounds surprisingly good, a song a Travis Scott fan wouldn't think twice about if it randomly showed up on their Spotify or SoundCloud weekly playlist. Not until they put a close ear to the lyrics, that is, which you can find over at the song's Genius lyric page.