Summary: A pioneering AI system successfully identifies violations of social norms. Using GPT-3, zero stroke text classification and automatic rule discovery, the system classifies social emotions into ten main types. He analyzes the written situations and determines with precision if they are positive or negative according to these categories.
This first study offers promising evidence that the approach can be broadened to encompass more social norms.
Highlights:
- The AI system uses ten categories of social emotions to identify violations of social norms.
- The system was tested on two large datasets of short texts, validating its models.
- This preliminary work, funded by DARPA, is considered an important step in improving cross-cultural language understanding and situational awareness.
Source: Ben Gurion University of the Negev
A researcher at Ben-Gurion University of the Negev has designed an AI system that identifies violations of social norms.
The DARPA-funded project is one of the first to tackle the automatic identification of social norm violations. While many social norms exist around the world, the violation of social norms boils down to a few general categories.
Professor Yair Neuman and his engineer Yochai Cohen built the system using GPT-3, zero-shot text classification and automatic rule discovery. The system used a binary of ten social emotions as categories.
DARPA commissioned the Computational Cultural Understanding (CCU) program to create cross-cultural language understanding technologies to improve the situational awareness and interactional effectiveness of a Department of Defense operator. Poor cross-cultural communication not only derails negotiations, but can also be a contributing factor leading to war, according to DARPA’s explanation of the program’s rationale.
Their findings were recently published in the prestigious journal Scientific reports.
Professor Neuman and his engineer trained the system to identify ten social emotions: competence, politeness, confidence, discipline, attention, friendliness, success, conformity, decency and loyalty. The system succeeded in characterizing a situation written under one of these ten classifiers and was able to perceive whether it was positive or negative.
The system was tested on two massive datasets of short texts and empirically proved the validity of the models.
“This is preliminary work, but it provides strong evidence that our approach is correct and can be extended to include more social norms,” says Professor Yair Neuman.
Professor Neuman is the head of the Functor Lab in the Department of Cognitive and Brain Sciences at BGU.
About this artificial intelligence research news
Author: Ehud Zion Waldoks
Source: Ben Gurion University of the Negev
Contact: Ehud Zion Waldoks – Ben Gurion University of the Negev
Picture: Image is credited to Neuroscience News
Original research: Free access.
“AI to identify violation of social norms” by Yair Neuman et al. Scientific reports
Abstract
AI to identify violation of social norms
Identifying social norms and their violation is a challenge faced by many computer science projects. This article presents a new approach to identify violations of social norms.
We used GPT-3, zero-hit classification, and automatic rule discovery to develop simple predictive models based on psychological insights.
Tested on two massive datasets, the models exhibit significant predictive performance and show that even complex social situations can be functionally analyzed using modern computational tools.
