Home » PyTorch, Mark Zuckerberg opens up to the world of free software

PyTorch, Mark Zuckerberg opens up to the world of free software

by admin
PyTorch, Mark Zuckerberg opens up to the world of free software

Artificial intelligence is among us to stay there. After a 70-year ride, the discipline born from the studies of Marvin Minsky, Claude Shannon and John McCarthy – who gave it its name in 1956 – has now translated into applied research, products and the market. The latest example of this long ride comes from Zuckerberg’s launch of the PyTorch Foundation “to accelerate advances in artificial intelligence (AI) research.” The foundation, wanted by the board of Meta-Facebook, however, will have a board of directors extended to representatives of both Meta and AMD, Amazon Web Services, as well as Google Cloud, Microsoft Azure and Nvidia. The project will be part of the Linux Foundation, a non-profit organization that supports the development of free software.

The project

Meta joins the Linux Foundation to accelerate AI advancements


PyTorch is an artificial intelligence software launched in 2016 for computer vision applications and natural language processing and, led by Meta-Facebook, has given life to over 150 thousand projects, as Mark Zuckerberg points out in a post. But why a foundation? Santosh Janardhan, Meta’s Vice President of Infrastructure explains: “The creation of the foundation ensures that decisions in the coming years are made in an open and transparent way by a diverse group of committee members. One of its top priorities will be to clearly separate the corporate governance from the technical one of PyTorch “.

European Union

Dominant position in search engines, confirmed the fine of over 4 billion for Google

See also  Electric vehicles for Europe - who are the losers?


The purpose of the foundation will therefore be to promote the adoption of tools for artificial intelligence by fostering and supporting an ecosystem of open source and vendor neutral projects and by making the most advanced tools, libraries and other AI components more accessible, but also by supporting the software through conferences and training courses.

Therefore, as reiterated by Janardhan, the foundation will undertake to respect four principles: a) remain open, b) maintain its neutral nature, c) guarantee fairness and d) shape a strong technical identity.

These are no small commitments, both for the aggressiveness of market forces and for the war and political interests that artificial intelligence tickles. On the other hand, both Vladimir Putin and Elon Musk, now on opposite sides of the Russian-Ukrainian conflict, have already expressed their thoughts: whoever dominates artificial intelligence will dominate the world.

Already today, artificial intelligence is at the basis of small and large daily technological innovations: from the search for the best price on the web to autonomous bus driving experiments in Estonia, up to the creation of artistic works using Dall-E software. But the most relevant sectors will certainly be that of AI-led air transport and unfortunately autonomous weapons, with respect to which scientists have already expressed their opposition.

But artificial intelligence with its computing power should also help us make our crops and cities more efficient, energy consumption and predict extreme weather events, lighten the work of legal assistants in finding sentences, automate strenuous work and dangerous to humans. A power that in the future will make accountants, insurance experts, clerks, secretaries and cashiers disappear. Not to mention call centers and taxi drivers or journalists who are already beginning to be replaced by AI.

See also  Fashion offers at Otto: Up to 70% discount on trendy shirts

As reported by the Ansa agency, however, Meta’s belief is that “Open science is the basis of our work in the field of AI. We believe that this approach accelerates progress in the development and distribution of new systems that will respond to real needs. and fundamental questions about the nature of artificial intelligence “. A scientific field open to both speculation and social progress where the ethical component of AI will be a topic of debate in the years to come.

The very choice of creating a foundation and putting it under the Linux Foundation’s hat has made more than one activist turn up their noses. Father Paolo Benanti, Professor of Ethics of Technologies and Artificial Intelligence at the Gregorian University of Rome, when asked about this, told us: “PyTorch represents the conjunction between a product that has industrial implications and serves to grind profits with something very useful for the scientific world which is “open” by definition. But it hides, like other similar products, an unresolved tension: the responsibility of software products. By placing it under the umbrella of the Linux Foundation, the PyTorch foundation will be open to an open community that will benefit, but reduce industrial responsibility. In addition, it is good to remember that opening up to the free software community is easy to scout for those intelligences that are worth even more than the data used to make machines work ”.

For those more nerdy it is possible to make a clarification: most of the deep learning applications are already written in PyTorch and in two other main frameworks: TensorFlow and JAX. The first, released in 2015, is a vast and reliable framework developed by the Google Brain team and runs many Google products, from speech recognition to email to Google Photos. M also JAX is a deep learning framework created, maintained and used by Google, but it is not a Google product. PyTorch now presents itself as its closest competitor. Already several pieces of deep learning software are built with PyTorch, including Tesla’s autopilot, Uber’s Pyro, and several others.

See also  Mima Ito loses in the Japanese trials and basically misses the women's singles quota at the Hangzhou Asian Games_Match_Chinese Player_Women's Doubles

You may also like

Leave a Comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.

This website uses cookies to improve your experience. We'll assume you're ok with this, but you can opt-out if you wish. Accept Read More

Privacy & Cookies Policy