Believable proxies of human behavior can empower interactive applications ranging from immersive environments to rehearsal spaces for interpersonal communication to prototyping tools. In this paper, we introduce generative agents–computational software agents that simulate believable human behavior. Generative agents wake up, cook breakfast, and head to work; artists paint, while authors write; they form opinions, notice each other, and initiate conversations; they remember and reflect on days past as they plan the next day. To enable generative agents, we describe an architecture that extends a large language model to store a complete record of the agent’s experiences using natural language, synthesize those memories over time into higher-level reflections, and retrieve them dynamically to plan behavior. We instantiate generative agents to populate an interactive sandbox environment inspired by The Sims, where end users can interact with a small town of twenty five agents using natural language. In an evaluation, these generative agents produce believable individual and emergent social behaviors: for example, starting with only a single user-specified notion that one agent wants to throw a Valentine’s Day party, the agents autonomously spread invitations to the party over the next two days, make new acquaintances, ask each other out on dates to the party, and coordinate to show up for the party together at the right time. We demonstrate through ablation that the components of our agent architecture–observation, planning, and reflection–each contribute critically to the believability of agent behavior. By fusing large language models with computational, interactive agents, this work introduces architectural and interaction patterns for enabling believable simulations of human behavior. Read More

#human

Nine brains, blue blood, instant camouflage: It’s no surprise that octopuses capture our interest and our imaginations. Science-fiction creators, in particular, have been inspired by these tentacled creatures.

An octopus’s remarkable intelligence makes it a unique subject for marine biologists and neuroscientists as well. Research has revealed the brain power of the octopus allows it to unscrew a jar or navigate a maze. But, like many children, the octopus also develops an impish tendency to push the boundaries of behavior. Several aquariums have found octopuses memorizing guard schedules to sneak into nearby tanks to steal fish; meanwhile, marine biologists have discovered that wild octopuses will punch fish… for no apparent reason.

According to Dr. Jennifer Maher, a professor at the University of Lethbridge in Canada, there are a “number of [different] types of learning [for octopuses]: cognitive tasks like tool use, memory of complex operations for future use, and observational learning.”

How does the distinct structure of the octopus’s brain enable all this complex behavior? No one had successfully studied wild or freely moving octopuses’ brain waves until a new study by researchers at the University of Naples Federico II in Italy and the Okinawa Institute of Science and Technology (OIST) in Japan, among others. In their Current Biology paper, the researchers tracked and monitored three captive but freely moving octopuses, analyzing their brain waves for the first time. Using recording electrodes, the researchers found a type of brain wave never before seen, along with brain waves that may be similar to some seen in human brains, possibly providing hints about the evolution of intelligence. Read More