Choose an application

• Reference Manager
• EndNote
• RefWorks (Direct export to RefWorks)

The objective of this master thesis is to find transfer learning
methods which can be applied in reinforcement learning problems.
This method has to improves the learning speed of an agent on a
target task, given a set of already trained policies on similar
simpler tasks. We used Q learning algorithm combined to neural
network to represent Q function that determine how an agent act
on it’s assigned task.
Using source network trained on simpler problems, we manage to
find an architecture that combines networks with an interface to
transfer knowledge in the new network. This technique allowed us
to solve the OpenAI Gym mountain car problem using source
networks trained on a smaller mountain and a faster car.
We tested more complex connection architectures to improve the
representation complexity of the combined networks.
We also implemented a method to learn multiple networks in
parallel to exploit the ensemble properties when learning a new
task. This method improves the convergence and results.
Finally we tested our implementation on the OpenAI Gym
Cartpole environment with succes.

transfer --- learning --- mountain --- car --- cartpole --- neural --- network --- concatenation --- merge --- openai --- gym --- reinforcement --- Ingénierie, informatique & technologie > Sciences informatiques

Choose an application

• Reference Manager
• EndNote
• RefWorks (Direct export to RefWorks)

"An electrifying vision of how we can safeguard AI's future for the public goodArtificial intelligence may be the most transformative technology of our time. As AI's power grows, so does the need to figure out what-and who-this technology is really for. AI Needs You argues that it is critical for society to take the lead in answering this urgent question and ensuring that AI fulfills its promise.Verity Harding draws inspiring lessons from the histories of three twentieth-century tech revolutions-the space race, in vitro fertilization, and the internet-to empower each of us to join the conversation about AI and its possible futures. Sharing her perspective as a leading insider in technology and politics, she rejects the dominant narrative, which often likens AI's advent to that of the atomic bomb. History points the way to an achievable future in which democratically determined values guide AI to be peaceful in its intent; to embrace limitations; serve purpose, not profit; and to be firmly rooted in societal trust. AI Needs You gives us hope that we, the people, can imbue AI with a deep intentionality that reflects our best values, ideals, and interests, and that serves the public good. AI will permeate our lives in unforeseeable ways, but it is clear that the shape of AI's future-and of our own-cannot be left only to those building it. It is up to us to guide this technology away from our worst fears toward a future that we can trust and believe in"-- "This brief and accessible book draws lessons from the history and governance of three recent transformative technologies - the space race, in-vitro fertilization (IVF), and the internet - to argue that society can and should take an active role in shaping the future of artificial intelligence (AI). It is a manifesto aimed at empowering the reader to participate in the conversations and political and democratic processes that will determine the intentions of AI, what values and regulations should guide its development, and its future. Artificial intelligence affects most of us every day, from determining what news we read and music we listen to, to influencing credit scores and legal decisions. As computers become smarter, and the amount of data available to train sophisticated algorithms grows, these unprecedented abilities are coming to play an ever more central role in how our society functions. Yet, as important as this new technology is and will be, it is little understood outside of an unaccountable, insular, and hard-to-scrutinize tech community. Creators of AI-driven systems often operate quickly, at large scale, and without any clear sense of societal purpose or understanding of the diversity and complexity of the human condition. As these systems become increasingly powerful, figuring out what, and who, this technology is for is critical, if we want to ensure that advances in AI will advance us as a species, inspire us as a people, and support the delicate fabric of society. What values framework permeates this new technology, which will permeate our lives? What values are needed - to guide us into a future of which we can all be proud? By looking back, Harding shows that science and technology are clearly not neutral, but inherently political, dictated by the human values and preferences of their time. Recognizing this gives us cause for hope; democratic and political judgement can and must be applied to today's scientific breakthroughs, to ensure that the values we all hold dear - not just the values of a few - will guide us into the new frontier. In four thematically structured chapters, in which Harding connects her historical examples to current considerations about AI, the author argues that AI should be peaceful in its intent, embrace limitations, serve purpose not profit, and be rooted in societal trust. History tells us that we can imbue AI with a deep intentionality that aligns with our best and brightest ideals, interests, and values and that serves the public good - but, to make this happen, the public must take part in this conversation"--

Technological innovations --- Technology and civilization. --- Technology and state. --- Artificial intelligence --- AI Bill of Rights. --- AI Needs You: How We Can Change AI's Future and Save Our Own. --- AI education: AI public good. --- AI ethics. --- AI fear: Artificial intelligence. --- AI governance. --- AI regulation. --- AI’s power grows. --- Artificial intelligence. --- Atomic Bomb. --- Bing. --- ChatGPT. --- Chatbot. --- Cold War. --- DeepMind. --- Embryo research/human embryology. --- Google. --- IVF (in vitro fertilization). --- Large Language Models (LLMs). --- Louise Joy Brown. --- Online security. --- OpenAI. --- Oppenheimer. --- Roe V. Wade. --- Silicon Valley. --- Space Race. --- Times 100 AI. --- United Nations Outer Space Treaty 1967. --- Warnock Commission. --- better education. --- better society. --- biotech. --- chatgpt. --- critical. --- democracy geopolitics. --- future of AI. --- generative ai. --- harmful. --- history of science. --- history. --- how can AI help: AI policy. --- how we can safeguard AI’s future for the public good. --- internet. --- life sciences. --- politics. --- safe AI: is AI good: is AI bad. --- satellites. --- science. --- societal trust. --- society. --- space race. --- technology. --- transformative technology. --- urgent: social responsibility. --- verity harding. --- Social aspects.