Welcoming @chrismeserole as the first Executive Director of the Frontier Model Forum:
AGI
-
What Algorithms Can Transformers Learn Study
By
–
What Algorithms can Transformers Learn? A Study in Length Generalization Zhou et al.: https://
arxiv.org/abs/2310.16028 #ArtificialIntelligence #DeepLearning #MachineLearning -
Accelerated Timeline Debate in AI Development Strategy
By
–
Would love to see an accelerated timeline, but might be quite a bit longer if we assume they're going to bide their time as usual.
-
Multi-Agent Simulation Framework for Strategy Testing
By
–
Let's recap. Agent: You describe your agents, their objectives and goals Environment: Define the environment using just text Simulation: Start the simulation and observe agents' strategies, gain valuable insights in a simulated environment
-
Simulating Agent Strategies in Controlled Environments
By
–
Simulation Now it's time to start the simulation, we select the environment, we add the agents we want to participate and start running the simulation. The simulation observes agents' strategies, gaining insights without real-world testing.
-
Creating Realistic Restaurant Environment Simulation
By
–
Create your Environment We create a simple Restaurant, in our case a Chinese restaurant. With a nice vibe and quite busy with loads of people inside.
-
Creating AI Agents with Objectives and Goals
By
–
Agents In the demo above we created a Chef and an Apprentice. You describe your agents, their objectives and goals using just text.
-
Language Models Discovering World Models for Biology Applications
By
–
“With natural language, there’s a developing view that these models may be able to discover world models and simulate aspects of the world when trained on enough data. In biology, i think there’s an exciting potential for this.” @alexrives speaking at the 2023 HAI Fall Conference
-
Designing AI Intrinsic Objectives to Prevent Sociopathic Behavior
By
–
You are right.
You can't control if a human is a sociopath.
But we can make sure an AI system won't be a sociopath because we get to design their intrinsic objectives. -
AI Alignment Easier Than Aligning Humans Through Design
By
–
No.
But I'm arguing that
a) we are familiar with the concept of aligning intelligent entities, whether they are children, adults, or corporations (which are super-human).
b) it is *much* easier to align AI than humans because we get to *design* the intrinsic motivations of AI