It's hard to scale UT in terms of param sharing because of the shared params across all layers (similar to ALBERT). If you want to have like a 1B UT model, it's going to be super slow. That said, we didn't try scaling non-shared UT which could be okay. More about
RESEARCH
-
Managing Generative AI Risks: Ethics and Safety
By
–
Managing the Risks of Generative AI https://
hbr.org/2023/06/managi
ng-the-risks-of-generative-ai
… #GenAI #ChatGPT #BingAI #ChatGPT #opensource #EthicalAI #Python #tech #chatgpt4 #AI #ML #AIEthics #chatgpt3 #code #GPT3 #GPT4 #GPT4AI #aichat #chatbot #ChatbotAI #BardAI #chatgpt5 #LLMs #LLM #GenerativeAI #GoogleBard -
Vatican Releases Comprehensive AI Ethics Handbook Guide
By
–
The #Vatican Releases Its Own AI Ethics Handbook! https://
gizmodo.com/pope-francis-v
atican-releases-ai-ethics-1850583076?utm_medium=sharefromsite&utm_source=gizmodo_twitter
… via @gizmodo #Robotics #robot #EthicalAI #OpenAI #bots #KillerRobot #roboticsainews #opensource #Python #tech #technology #cobot #humanoid #AI #ML #Algorithms #AIEthics #Robotech #RobotDesign #Engineering -
AI Safety Discourse: Moving Beyond Doomer-Denialist Polarization
By
–
The ongoing discourse surrounding AI safety and its potential for existential risks seems to have escalated into an intense polarization, with participants taking entrenched positions – "AI Doomers" versus "AI Denialists." Such adversarial classification is not helpful for a
-
Why ACT Comes Before Action: AI Ethics and Thought Leadership
By
–
There is a reason ACT starts with "A" not "T". Thank you for the food for thought @JimMarous and @JamesClear
-
ML absorbs AI engineering through test-time to training-time shift
By
–
I think ML gradually eats AI engineering as that which is done at test time gets absorbed into training. Kind of similar to relationship between AI and HCI.
-
Earth’s Systematic Approach: Banning Everything Except the Most Dangerous
By
–
See, it's not that Earth is systematically too cautious. It's that Earth is systematically always wrong. They'll ban everything *except* gain-of-function research in bio and AI. They won't let that stopped clock be right even once a day.
-
OpenReview Enhances Transparency in Expert Reviewer Identification Process
By
–
Partly to address that last point, the openreview page of these papers lists out explicitly who is the author that is an Expert Reviewer.
-
Top 10 AI Companies to Watch in 2023
By
–
Top 10 AI Companies to Watch in 2023: The quest for the best AI companies heightens in the super-evolving terrain of artificial intelligence. The need for cutting-edge technology and ground-breaking innovation has risen dramatically as we progress… https://
analyticsvidhya.com/blog/2023/05/t
op-ai-companies/?utm_source=dlvr.it&utm_medium=twitter
…