Researchers propose bias fix for GPT-3 and other language models

Few-shot learning, or the ability to learn tasks from a few examples, is a key aspect of human intelligence. Large AI natural language models like OpenAI’s GPT-3 can perform few-shot learning without fine-tuning. But despite the promise of few-shot learning, new research finds that the accuracy of language models — particularly GPT-3 — can be […]

Facebook researchers propose ‘pre-fine-tuning’ to improve language model performance

Machine learning researchers have achieved remarkable success with language model pretraining, which uses self-supervision, a training technique that doesn’t require labeled data. Pretraining refers to training a model with one task to help it recognize patterns that can be applied to a range of other tasks. In this way, pretraining imitates the way human beings […]

Researchers propose LEAF, a frontend for developing AI classification algorithms

In machine learning, mel-filterbanks — fixed, hand-engineered representations of sound — are often used to train algorithms that classify sound. Decades after the design of mel-filterbanks, research shows that they exhibit desirable mathematical properties for representation learning; in other words, they represent strong audio feature. But the design of mel-filterbanks is also flawed by biases, […]

Computer vision researchers develop city-scale smart tracking platform

Surveillance cameras already dot or blanket major cities, but it’s not necessarily easy to continuously track a person or object moving through multiple locations and cameras. Now researchers from the Indian Institute of Science are applying advanced artificial intelligence algorithms to this complex challenge, creating a software platform called Anveshak — Hindi for “investigator” — that […]

Uber researchers propose AI language model that emphasizes positive and polite responses

AI-powered assistants like Siri, Cortana, Alexa, and Google Assistant are pervasive. But for these assistants to engage users and help them to achieve their goals, they need to exhibit appropriate social behavior and provide informative replies. Studies show that users respond better to social language in the sense that they’re more responsive and likelier to […]

Few researchers thoroughly test adversarial attacks on surveillance AI in the real world

Advances in machine learning, particularly in computer vision and facial recognition, have enabled systems deployed by governments and private corporations to erode expectations of privacy. Adversarial attacks, or techniques that attempt to fool models with deceptive images, videos, and text, are one way to combat surveillance. But in a new paper, researchers from Microsoft, Harvard, […]

Researchers propose ‘safe’ reinforcement learning algorithm for dangerous scenarios

Researchers have proposed a method for allowing reinforcement learning algorithms to accumulate knowledge while erring on the side of caution. The team, which hails from the University of Toronto, the Vector Institute, and the University of California, Berkeley,  claims this approach can achieve competitive performance while incurring lower catastrophic failure rates during training compared to […]

Researchers find evidence of racial, gender, and socioeconomic bias in chest X-ray classifiers

Google and startups like Qure.ai, Aidoc, and DarwinAI are developing AI and machine learning systems that classify chest X-rays to help identify conditions like fractures and collapsed lungs. Several hospitals, including Mount Sinai, have piloted computer vision algorithms that analyze scans from patients with the novel coronavirus. But research from the University of Toronto, the […]

Researchers find cutting-edge language models fall short in basic reasoning

Even sophisticated language models such as OpenAI’s GPT-3 struggle with socially important topics like morality, history, and law. That’s the top-line finding from a new paper coauthored by Columbia, University of Chicago, and University of California, Berkeley researchers that proposes a 57-task test to measure models’ ability to reason. Models must possess problem-solving abilities and […]

Microsoft researchers use Bing data to track shifting human needs during the pandemic

Research about the pandemic has mostly focused on epidemiology, but health data alone can’t inform policies — decisions should ideally take into account the impacts on communities. Driven by this maxim, Microsoft researchers developed a framework to characterize the changes in people’s physiological, socioeconomic, and psychological needs during the pandemic. After applying the framework to […]