A Privacy Threat Revealed: How researchers cracked InstaHide for computer vision.
Vision

A Privacy Threat Revealed: How researchers cracked InstaHide for computer vision.

With access to a trained model, an attacker can use a reconstruction attack to approximate its training data. A method called InstaHide recently won acclaim for promising to make such examples unrecognizable to human eyes while retaining their utility for training.

February 3, 20212 min read
Every Picture Tells a Story: Facebook expands automated alternative text.
Vision

Every Picture Tells a Story: Facebook expands automated alternative text.

Facebook expanded a system of vision, language, and speech models designed to open the social network to users who are visually impaired. A Facebook service that describes photos in a synthesized voice now recognizes 1,200 visual concepts — 10 times more than the previous version.

January 27, 20212 min read
Every Picture Tells a Story: Facebook expands automated alternative text.
Vision

Every Picture Tells a Story: Facebook expands automated alternative text.

Facebook expanded a system of vision, language, and speech models designed to open the social network to users who are visually impaired. A Facebook service that describes photos in a synthesized voice now recognizes 1,200 visual concepts — 10 times more than the previous version.

January 27, 20212 min read
Draw a Gun, Trigger an Algorithm: These AI-enabled security cameras automatically ID guns.
Vision

Draw a Gun, Trigger an Algorithm: These AI-enabled security cameras automatically ID guns.

Computer vision is alerting authorities the moment someone draws a gun. Several companies offer deep learning systems that enable surveillance cameras to spot firearms and quickly notify security guards or police.

January 20, 20211 min read
Adversarial Helper: Adversarial learning can improve vision and NLP.
Vision

Adversarial Helper: Adversarial learning can improve vision and NLP.

Models that learn relationships between images and words are gaining a higher profile. New research shows that adversarial learning, usually a way to make models robust to deliberately misleading inputs, can boost vision-and-language performance.

January 20, 20212 min read
Adversarial Helper: Adversarial learning can improve vision and NLP.
Vision

Adversarial Helper: Adversarial learning can improve vision and NLP.

Models that learn relationships between images and words are gaining a higher profile. New research shows that adversarial learning, usually a way to make models robust to deliberately misleading inputs, can boost vision-and-language performance.

January 20, 20212 min read
Draw a Gun, Trigger an Algorithm: These AI-enabled security cameras automatically ID guns.
Vision

Draw a Gun, Trigger an Algorithm: These AI-enabled security cameras automatically ID guns.

Computer vision is alerting authorities the moment someone draws a gun. Several companies offer deep learning systems that enable surveillance cameras to spot firearms and quickly notify security guards or police.

January 20, 20211 min read
Tell Me a Picture: OpenAI's two new multimodal AI models, CLIP and DALL·E
Vision

Tell Me a Picture: OpenAI's two new multimodal AI models, CLIP and DALL·E

Two new models show a surprisingly sharp sense of the relationship between words and images. OpenAI, the for-profit research lab, announced a pair of models that have produced impressive results in multimodal learning: DALL·E.

January 13, 20212 min read
The Fax About Tracking Covid: A deep learning system for sorting critical Covid-19 cases.
Vision

The Fax About Tracking Covid: A deep learning system for sorting critical Covid-19 cases.

A pair of neural networks is helping to prioritize Covid-19 cases for contact tracing. The public health department of California’s Contra Costa County is using deep learning to sort Covid-19 cases reported via the pre-internet technology known as fax.

January 13, 20212 min read
Tell Me a Picture: OpenAI's two new multimodal AI models, CLIP and DALL·E
Vision

Tell Me a Picture: OpenAI's two new multimodal AI models, CLIP and DALL·E

Two new models show a surprisingly sharp sense of the relationship between words and images. OpenAI, the for-profit research lab, announced a pair of models that have produced impressive results in multimodal learning: DALL·E.

January 13, 20212 min read
How Much For That Vintage Gucci?: An AI system that appraises luxury handbags
Vision

How Much For That Vintage Gucci?: An AI system that appraises luxury handbags

Computer vision is helping people resell their used designer handbags. Rebag, a resale company for luxury handbags, watches, and jewelry, launched Clair AI, an app that automatically appraises second-hand bags from brands like Gucci, Hermes, and Prada.

January 13, 20211 min read
How Much For That Vintage Gucci?: An AI system that appraises luxury handbags
Vision

How Much For That Vintage Gucci?: An AI system that appraises luxury handbags

Computer vision is helping people resell their used designer handbags. Rebag, a resale company for luxury handbags, watches, and jewelry, launched Clair AI, an app that automatically appraises second-hand bags from brands like Gucci, Hermes, and Prada.

January 13, 20211 min read
The Fax About Tracking Covid: A deep learning system for sorting critical Covid-19 cases.
Vision

The Fax About Tracking Covid: A deep learning system for sorting critical Covid-19 cases.

A pair of neural networks is helping to prioritize Covid-19 cases for contact tracing. The public health department of California’s Contra Costa County is using deep learning to sort Covid-19 cases reported via the pre-internet technology known as fax.

January 13, 20212 min read
Representing the Underrepresented: Many important AI datasets contain bias.
Vision

Representing the Underrepresented: Many important AI datasets contain bias.

Some of deep learning’s bedrock datasets came under scrutiny as researchers combed them for built-in biases. Researchers found that popular datasets impart biases against socially marginalized groups to trained models due to the ways the datasets were compiled, labeled, and used.

December 23, 20202 min read
Representing the Underrepresented: Many important AI datasets contain bias.
Vision

Representing the Underrepresented: Many important AI datasets contain bias.

Some of deep learning’s bedrock datasets came under scrutiny as researchers combed them for built-in biases. Researchers found that popular datasets impart biases against socially marginalized groups to trained models due to the ways the datasets were compiled, labeled, and used.

December 23, 20202 min read

Subscribe to The Batch

Stay updated with weekly AI News and Insights delivered to your inbox