Goto

Collaborating Authors

Results


A Distributional Approach to Controlled Text Generation

arXiv.org Artificial Intelligence

We propose a Distributional Approach to address Controlled Text Generation from pre-trained Language Models (LMs). This view permits to define, in a single formal framework, "pointwise" and "distributional" constraints over the target LM -- to our knowledge, this is the first approach with such generality -- while minimizing KL divergence with the initial LM distribution. The optimal target distribution is then uniquely determined as an explicit EBM (Energy-Based Model) representation. From that optimal representation we then train the target controlled autoregressive LM through an adaptive distributional variant of Policy Gradient. We conduct a first set of experiments over pointwise constraints showing the advantages of our approach over a set of baselines, in terms of obtaining a controlled LM balancing constraint satisfaction with divergence from the initial LM (GPT-2). We then perform experiments over distributional constraints, a unique feature of our approach, demonstrating its potential as a remedy to the problem of Bias in Language Models. Through an ablation study we show the effectiveness of our adaptive technique for obtaining faster convergence.


SYNTH- Changing the Perspectives of Audience towards AI through Cinema

#artificialintelligence

In a recent Podcast series with Analytics Insight, Lara Heller, British Filmmaker, Actress and Producer shares her views on how disruptive technologies are transforming everyday lives and impacting the film industry. Lara is of producing and acting in an upcoming movie SYNTH that takes us into the world of Artificial Intelligence. Having her roots in England, she is half German and half Persian. Lara plays the lead role as well produced SYNTH that explores the possibilities of AI. The name SYNTH is derived from the word synthetic and is inspired by the true story of an AI application called Sophia, a Humanoid Robot made by Hanson Robotics in Hong Kong.


Introducing SYNTH - The True Story of A Humanoid with British Actress Lara Heller

#artificialintelligence

Let's dive into the realms of Artificial Intelligence. We're thrilled to announce our latest podcast with Lara Heller, British Filmmaker, Actress and Producer of an upcoming movie SYNTH. SYNTH โ€“ Winner of Venice Film Awards for Best Sci-Fi is inspired by the true story of an AI application called Sophia, a Humanoid Robot made by Hanson Robotics in Hong Kong. Sophia is the first AI in the world to receive citizenship in Saudi Arabia. She further shares her views on how disruptive technologies are transforming everyday lives and impacting the film industry.


AutoETER: Automated Entity Type Representation for Knowledge Graph Embedding

arXiv.org Artificial Intelligence

Recent advances in Knowledge Graph Embedding (KGE) allow for representing entities and relations in continuous vector spaces. Some traditional KGE models leveraging additional type information can improve the representation of entities which however totally rely on the explicit types or neglect the diverse type representations specific to various relations. Besides, none of the existing methods is capable of inferring all the relation patterns of symmetry, inversion and composition as well as the complex properties of 1-N, N-1 and N-N relations, simultaneously. To explore the type information for any KG, we develop a novel KGE framework with Automated Entity TypE Representation (AutoETER), which learns the latent type embedding of each entity by regarding each relation as a translation operation between the types of two entities with a relation-aware projection mechanism. Particularly, our designed automated type representation learning mechanism is a pluggable module which can be easily incorporated with any KGE model. Besides, our approach could model and infer all the relation patterns and complex relations. Experiments on four datasets demonstrate the superior performance of our model compared to state-of-the-art baselines on link prediction tasks, and the visualization of type clustering provides clearly the explanation of type embeddings and verifies the effectiveness of our model.


Towards Conversational Recommendation over Multi-Type Dialogs

arXiv.org Artificial Intelligence

We propose a new task of conversational recommendation over multi-type dialogs, where the bots can proactively and naturally lead a conversation from a non-recommendation dialog (e.g., QA) to a recommendation dialog, taking into account user's interests and feedback. To facilitate the study of this task, we create a human-to-human Chinese dialog dataset \emph{DuRecDial} (about 10k dialogs, 156k utterances), which contains multiple sequential dialogs for every pair of a recommendation seeker (user) and a recommender (bot). In each dialog, the recommender proactively leads a multi-type dialog to approach recommendation targets and then makes multiple recommendations with rich interaction behavior. This dataset allows us to systematically investigate different parts of the overall problem, e.g., how to naturally lead a dialog, how to interact with users for recommendation. Finally we establish baseline results on DuRecDial for future studies. Dataset and codes are publicly available at https://github.com/PaddlePaddle/models/tree/develop/PaddleNLP/Research/ACL2020-DuRecDial.


Piracy app launches child-friendly mode amid lockdown streaming surge

The Independent - Tech

A popular app dubbed the "Netflix of piracy" has launched a child-friendly mode that filters out inappropriate content, as online pirates seek to profit from a recent surge in streaming. The new feature in Popcorn Time appears to be borrowed from similar parental controls on legitimate streaming services like Netflix and Amazon Prime Video. Much of the content hosted on Popcorn Time is also taken from its legal counterparts, offering users a free but illegal way to watch streams and torrents of films and TV shows. "The applications provide a free alternative to subscription-based video streaming services such as Netflix," the app states on its website. "Popcorn Time is constantly searching all over the web for the best torrents from the most important sites... If the movie or TV show is out there, Popcorn Time will find the best version possible and start streaming it right away."



AI In Gaming 2020 speaker interview: Andrew Pearson, Founder and MD, Intelligencia Limited - CalvinAyre.com

#artificialintelligence

Consistent in their quest to spearhead innovative, groundbreaking events, Eventus International is hosting the first ever AI In Gaming 2020 summit in Dubai on 26 and 27 February at Crowne Plaza Dubai. Joining a lineup of top international industry experts, is Andrew Pearson, founder and MD of Intelligencia Limited, who will be speaking at AI In Gaming 2020. Andrew Pearson was born in Pakistan, grew up in Singapore and was educated in England and America. With a degree in psychology from UCLA, Pearson has had a varied career in IT, marketing, mobile technology, social media and entertainment.In 2011, Pearson relocated to Hong Kong to open Qualex Asia Limited, bringing its parent company's experience into the ASEAN region. Pearson is the Managing Director of Intelligencia Limited, a leading implementer of BI, CI, data warehousing, data modeling, predictive analytics, data visualisation, digital marketing, mobile, social media and cloud solutions for the gaming, finance, telco, hospitality and retail industries.


In the Age of AI (full film) FRONTLINE

#artificialintelligence

A documentary exploring how artificial intelligence is changing life as we know it -- from jobs to privacy to a growing rivalry between the U.S. and China. FRONTLINE investigates the promise and perils of AI and automation, tracing a new industrial revolution that will reshape and disrupt our world, and allow the emergence of a surveillance society. This journalism is made possible by viewers like you. Support your local PBS station here: http://www.pbs.org/donate Find us on the PBS Video App where there are more than 250 FRONTLINE documentaries available for you to watch any time: https://to.pbs.org/FLVideoApp


Chinese Tech Titan Baidu Goes to BAT, Taps Hollywood And Madison Avenue

#artificialintelligence

One of China's tech titans, Baidu, is pursuing a new global business twist that takes the Beijing-based search company far from its roots in search to Hollywood movie promotion, Madison Avenue advertising and social media. It sounds far fetched, and the timing might be better given US-China tensions. But Baidu is making a play beyond China and is gaining traction with its latest outreach that blends search, artificial intelligence, advertising, and media overseas. Baidu has recently aligned with social networks Snap, Reddit and Pinterest, and movie producers Sony and Paramount, among others, in about a dozen sales partnerships. Baidu's distinct strategy for its global business unit is part of a trend of Chinese technology companies going outside China and transitioning from an original, core line of business.