Google to Urge to Solve Global Problems with AI

Companies selected as part of the AI ​​Impact Challenge will share a $ 25 million grant from the company
31 October 2018   802

Google urged non-profit, scientific and public organizations to suggest ways of using AI to solve social, humanitarian and environmental problems. Companies selected as part of the AI ​​Impact Challenge will share a $ 25 million grant from Google and will be assisted by AI specialists and will participate in the Launchpad Accelerator program. In the spring of 2019, an international team of experts will help Google choose the winning projects.

Google gave examples of successfully implemented projects that AI Impact Challenge participants can focus on:

  • Protection of Nature. Daniel de Leon used machine learning to analyze 100,000 hours of sounds made by rare whale species in the Pacific. Now AI automatically recognizes and classifies these sounds. In the future, scientists hope to use it to preserve rare mammals.
  • Fighting unemployment. The Harambee Youth Employment Accelerator project has helped more than 50,000 people in South Africa find jobs that do not require special skills.
  • Flood forecasting. Google developers have combined physical modeling and machine learning to predict floods.
  • Prevent forest fires. Two high school students from California created a device that uses AI to identify areas at risk of forest fires.
  • Baby health. The Canadian company Ubenwa has developed a mobile application that determines generic asphyxia by infant crying. Timely measures help reduce the risk of negative consequences for the newborn.

For those wishing to participate, but not sufficiently knowledgeable about machine learning, the company has compiled a manual.

When it comes to the use of AI, Google’s management is committed to ethical and reputational practices. In early October 2018, the company refused to participate in the Pentagon’s tender for $ 10 billion. A spokesman for the company said that this project may be contrary to the opinion of developers on the development of artificial intelligence.

MelNet Algorithm to Simulate Person's Voice

It analyzes the spectrograms of the audio tracks of the usual TED Talks, notes the speech characteristics of the speaker and reproduces short replicas
11 June 2019   339

Facebook AI Research team has developed a MelNet algorithm that synthesizes speech with characteristics specific to a particular person. For example, it learned to imitate the voice of Bill Gates.

MelNet analyzes the spectrograms of the audio tracks of the usual TED Talks, notes the speech characteristics of the speaker and reproduces short replicas.

Just the length of the replicas limits capabilities of the algorithm. It reproduces short phrases very close to the original. However, the person's intonation changes when he speaks on different topics, with different moods, different pitches. The algorithm is not yet able to imitate this, therefore long sentences sound artificially.

MIT Technology Review notes that even such an algorithm can greatly affect services like voice bots. There just all communication is reduced to an exchange of short remarks.

A similar approach - analysis of speech spectrograms - was used by scientists from Google AI when working on the Translatotron algorithm. This AI is able to translate phrases from one language to another, preserving the peculiarities of the speaker's speech.