Content deleted Content added
m →Participants: Task 14: cs1 template fixes: misused |publisher= (0×/1×); skipped: conflicting periodical (2×); ext text skip (1×); |
|||
Line 100:
=== GPT2 ===
GPT2 (2019) is an AI system that generates text matching its input in subject and tone. For example, when fed the first sentence of George Orwell's novel ''[[Nineteen Eighty-Four]]'' it produces plausible futuristic fiction set in China. Unlike previous OpenAI products, GPT2 has not been released to the public out of concerns of potential misuse, including applications for writing [[fake news]].<ref>{{Cite paper|last = Hern|first = Alex| title = New AI fake text generator may be too dangerous to release, say creators| newspaper = The Guardian| date = 14 February 2019|url = https://www.theguardian.com/technology/2019/feb/14/elon-musk-backed-ai-writes-convincing-news-fiction| access-date = 14 February 2019}}</ref> Much of the academic community is skeptical that GPT2 poses a significant threat. The [[Allen Institute for Artificial Intelligence]] followed up with a tool to detect "neural fake news".<ref>{{cite news |last1=Schwartz |first1=Oscar |title=Could ‘fake text’ be the next global political threat? |url=https://www.theguardian.com/technology/2019/jul/04/ai-fake-text-gpt-2-concerns-false-information |accessdate=16 July 2019 |work=The Guardian |date=4 July 2019}}</ref> Other researchers, like Jeremy Howard, warn of "the technology to totally fill Twitter, email, and the web up with reasonable-sounding, context-appropriate prose, which would drown out all other speech and be impossible to filter".<ref>{{cite news |last1=Vincent |first1=James |title=OpenAI’s new multitalented AI writes, translates, and slanders |url=https://www.theverge.com/2019/2/14/18224704/ai-machine-learning-language-models-read-write-openai-gpt2 |accessdate=16 July 2019 |work=The Verge |date=14 February 2019}}</ref>
==GYM Retro==
Gym Retro is a platform for reinforcement learning research on games. Gym Retro is used to conduct research on RL algorithms and study generalization. Prior research in RL has mostly focused on optimizing agents to solve single tasks. Gym Retro gives the ability to generalize between games with similar concepts but different appearances.
==See also==
| |||