Collaborating Authors


Top Big AI Trends and Challenges Impacting Media, Advertising & Entertainment Industry


I recently interviewed some of the top data science leaders from Comcast/Freewheel, Condé Nast, ViacomCBS, Audoir, USA Today Network, and Samba TV on the biggest trends, challenges, and opportunities they see for ML & AI in media, advertising, & entertainment -- and what the future may hold. What are some of the biggest trends you'll see being adopted by the entertainment and media industries? Christopher Whitely, Senior Director of Applied Analytics at Comcast/FreeWheel, shares "There are a few areas that we'll see adopted by M&E industries in the coming months and years, including more contextual advertising, where advertising creative assets are matched to appropriate program content algorithmically. Federated learning is also a new trend, which refers to modeling using machine learning without sharing data sets. Privacy is important, so I expect we'll see continued use of aggregated customer segments and clean rooms for marketing and analytics. Also, lookalike models will help advertisers reach potential customers and optimize campaigns for the greatest effect."

Disney made a demon Groot robot that dances, and sure why the hell not


Are you excited to walk around Disney's Avengers Campus, which opens at Disneyland California on June 4? No? There's a pandemic, you say? Let's try this another way: Would you be excited to walk around Disney's Avengers Campus if you knew there was a free-roaming robot modeled after Teen Groot that you might run into? Sure, wait until you're vaccinated. "Robot Teen Groot-as-tourist attraction" is still more "if" than "when" at this point -- the YouTube description makes that clear. But the Walt Disney Imagineering Research & Development team did indeed build the thing, and it could offer a glimpse at what some of the future character interactions at Disney parks might look like.

'Black Mirror' Episode Comes To Life With NYPD Robot Dogs, Internet Terrified

International Business Times

The internet is terrified of the New York Police Department's newest "canine" on unit: Digidog, a robo-dog that the Netflix series "Black Mirror" warned of. After a video went viral of Digidog in action, the internet started comparing it to Series 4 Episode 5, "Metalhead," where human society is no longer in existence and has been overrun by robot dogs. Some fear that this new invention could eventually turn into something negative. It was first deployed in February when men were being held hostage in a Bronx apartment and the robot was able to see how safe it was and if it was safe for the police to enter, the New York Times reported. The creators of Digidog, Boston Dynamics, explained that these devices won't be used as a weapon, but a political art collective has shared a few examples of how easy it is for things to go downhill fast, including a handful of Muslim Americans being killed by drones, according to the Guardian.

All NLP Tasks Are Generation Tasks: A General Pretraining Framework Artificial Intelligence

There have been various types of pretraining architectures including autoregressive models (e.g., GPT), autoencoding models (e.g., BERT), and encoder-decoder models (e.g., T5). On the other hand, NLP tasks are different in nature, with three main categories being classification, unconditional generation, and conditional generation. However, none of the pretraining frameworks performs the best for all tasks, which introduces inconvenience for model development and selection. We propose a novel pretraining framework GLM (General Language Model) to address this challenge. Compared to previous work, our architecture has three major benefits: (1) it performs well on classification, unconditional generation, and conditional generation tasks with one single pretrained model; (2) it outperforms BERT-like models on classification due to improved pretrain-finetune consistency; (3) it naturally handles variable-length blank filling which is crucial for many downstream tasks. Empirically, GLM substantially outperforms BERT on the SuperGLUE natural language understanding benchmark with the same amount of pre-training data. Moreover, GLM with 1.25x parameters of BERT-Large achieves the best performance in NLU, conditional and unconditional generation at the same time, which demonstrates its generalizability to different downstream tasks.

The Morning After: Netflix dominates Oscar nominations during a pandemic year


In news that probably won't shock you all that much, this year's Oscars reflect a year spent mostly indoors and not in movie theaters. The Academy has announced the nominees for the 2021 Oscars, and Netflix is, again, the frontrunner, grabbing 31 nominations. All those nominations won't guarantee wins, sure, but David Fincher's Mank dominated the shortlist. Its 10 nominations included Best Picture, Best Director, Best Actor (Gary Oldman) and Best Supporting Actress (Amanda Seyfried). Amazon picked up nominations for Borat: Subsequent Moviefilm, and Hulu's The United States vs. Billie Holiday was also recognized.

The Morning After: 'Cyberpunk 2077' runs into another delay


Cyberpunk 2077's woes have continued long after the game launched, with all the issues that entailed. CD Projekt Red announced yesterday that we'll have to wait until the second half of March for the next big patch. The developer cited that recent ransomware hack as the major culprit -- it initially planned to launch the 1.2 patch in February. As you're probably aware, February ends this week. The news is especially frustrating for PS5 owners as the game hasn't returned to the PlayStation Store since it was pulled.

Here are 5 Mars-themed movies worth watching tonight

USATODAY - Tech Top Stories

Purchases you make through our links may earn us a commission. Countless Americans watched with awe as NASA's latest robotic explorer, the Perseverance rover, landed safely on Mars earlier today. While you might not be on a mission to Mars personally, you can still make a sojourn to the Red Planet without ever ditching your sweats (or more importantly, leaving the couch), courtesy of these five movies streaming now on major platforms like Hulu and Disney . From green-suited space invaders firing ray guns to survival sagas pitting man against an inhospitable wilderness, these movies about Mars are an absolute must-watch for anyone who's still got space travel on the mind tonight. If Martians actually landed on Earth, would they be friends... or foes?

Knowledge Generation -- Variational Bayes on Knowledge Graphs Artificial Intelligence

This thesis is a proof of concept for the potential of Variational Auto-Encoder (VAE) on representation learning of real-world Knowledge Graphs (KG). Inspired by successful approaches to the generation of molecular graphs, we evaluate the capabilities of our model, the Relational Graph Variational Auto-Encoder (RGVAE). The impact of the modular hyperparameter choices, encoding through graph convolutions, graph matching and latent space prior, is compared. The RGVAE is first evaluated on link prediction. The mean reciprocal rank (MRR) scores on the two datasets FB15K-237 and WN18RR are compared to the embedding-based model DistMult. A variational DistMult and a RGVAE without latent space prior constraint are implemented as control models. The results show that between different settings, the RGVAE with relaxed latent space, scores highest on both datasets, yet does not outperform the DistMult. Further, we investigate the latent space in a twofold experiment: first, linear interpolation between the latent representation of two triples, then the exploration of each latent dimension in a $95\%$ confidence interval. Both interpolations show that the RGVAE learns to reconstruct the adjacency matrix but fails to disentangle. For the last experiment we introduce a new validation method for the FB15K-237 data set. The relation type-constrains of generated triples are filtered and matched with entity types. The observed rate of valid generated triples is insignificantly higher than the random threshold. All generated and valid triples are unseen. A comparison between different latent space priors, using the $\delta$-VAE method, reveals a decoder collapse. Finally we analyze the limiting factors of our approach compared to molecule generation and propose solutions for the decoder collapse and successful representation learning of multi-relational KGs.

The Morning After: LG might get out of the smartphone business


In the US, today is Inauguration Day, and as Joe Biden prepares to take the oath as our 46th president, it's worth taking a look back at the discussions four years ago. Back then, the "most tech-savvy" president exited as all eyes turned to Donald Trump trading in his Android Twitter machine for a secure device. We know how things went after that. Donald Trump isn't tweeting anymore (at least not from his main accounts), and the country is struggling through a pandemic. The outgoing president just saw his temporary YouTube ban extended and, in one of his last official acts, pardoned Anthony Levandowski for stealing self-driving car secrets from Google's subsidiary Waymo.

Podcast: Attention shoppers–you're being tracked

MIT Technology Review

In some stores, sophisticated systems are tracking customers in almost every imaginable way, from recognizing their faces to gauging their age, their mood, and virtually gussying them up with makeup. The systems rarely ask for people's permission, and for the most part they don't have to. In our season 1 finale, we look at the explosion of AI and face recognition technologies in retail spaces, and what it means for the future of shopping. This episode was reported and produced by Jennifer Strong, Anthony Green, Tate Ryan-Mosley, Emma Cillekens and Karen Hao. Strong: Retailers have been using face recognition and AI tracking technologies for years. And what if you could know about the presence of violent criminals before they act? With Face First you can stop crime before it starts.] It detects faces, voices, objects and claims it can analyze behavior. But face recognition systems have a well-documented history of misidentifying women and people of color. And they're trying to sell it and impose it on the entirety of the country?] Strong: This is Representative Alexandria Ocasio-Cortez at a 2019 congressional hearing on facial recognition.