Sign up to save your podcasts
Or
Share Will AI Ever Have Common Sense?
Share to email
Share to Facebook
Share to X
Share to email
Share to Facebook
Share to X
Or
Will AI Ever Have Common Sense?
This article from Quanta Magazine explores the current capabilities and limitations of large language models (LLMs) such as ChatGPT, examining whether these AIs can truly possess common sense.
Choi, who is a leading researcher in the field, explains how LLMs are trained, using a method called “backpropagation,” which involves adjusting the weights of a neural network to maximize the probability of predicting the correct word sequence. However, she argues that this process is very different from how humans learn. Humans, unlike LLMs, learn through curiosity, experimentation, and the ability to abstract concepts and understand the world through a combination of physical experience and social interaction.
While LLMs have made impressive strides in acquiring common sense, this knowledge is often fragile and limited, lacking the robustness of human understanding. Choi also touches on the ethical concerns related to the increasing power of LLMs and calls for more transparency in the data and methods used to train them. She emphasizes the need for a collective effort, involving governments, companies, and the public, to develop responsible guidelines for the development and deployment of these powerful technologies.
Source: https://www.quantamagazine.org/will-ai-ever-have-common-sense-20240718/
#AI #neuralnetworks #machinelearning #aiethics #artificialintelligence #computerscience #naturallanguageprocessing #thejoyofwhy #deepdivejournal
View all episodes
By Deep Dive Journal PodcastThis article from Quanta Magazine explores the current capabilities and limitations of large language models (LLMs) such as ChatGPT, examining whether these AIs can truly possess common sense.
Choi, who is a leading researcher in the field, explains how LLMs are trained, using a method called “backpropagation,” which involves adjusting the weights of a neural network to maximize the probability of predicting the correct word sequence. However, she argues that this process is very different from how humans learn. Humans, unlike LLMs, learn through curiosity, experimentation, and the ability to abstract concepts and understand the world through a combination of physical experience and social interaction.
While LLMs have made impressive strides in acquiring common sense, this knowledge is often fragile and limited, lacking the robustness of human understanding. Choi also touches on the ethical concerns related to the increasing power of LLMs and calls for more transparency in the data and methods used to train them. She emphasizes the need for a collective effort, involving governments, companies, and the public, to develop responsible guidelines for the development and deployment of these powerful technologies.
Source: https://www.quantamagazine.org/will-ai-ever-have-common-sense-20240718/
#AI #neuralnetworks #machinelearning #aiethics #artificialintelligence #computerscience #naturallanguageprocessing #thejoyofwhy #deepdivejournal