Face verification remains a challenging problem in very complex conditions with large variations such as pose, illumination, expression, and occlusions. This problem is exacerbated when we rely unrealistically on a single training data source, which is often insufficient to cover the intrinsically complex face variations. This paper proposes a principled multi-task learning approach based on Discriminative Gaussian Process Latent Variable Model, named GaussianFace, to enrich the diversity of training data. In comparison to existing methods, our model exploits additional data from multiple source-domains to improve the generalization performance of face verification in an unknown target-domain. Importantly, our model can adapt automatically to complex data distributions, and therefore can well capture complex face variations inherent in multiple sources. Extensive experiments demonstrate the effectiveness of the proposed model in learning from diverse data sources and generalize to unseen domain. Specifically, the accuracy of our algorithm achieves an impressive accuracy rate of 98.52% on the well-known and challenging Labeled Faces in the Wild (LFW) benchmark. For the first time, the human-level performance in face verification (97.53%) on LFW is surpassed.

Ref: Surpassing Human-Level Face Verification Performance on LFW with GaussianFace – Cornell University Lab

The online movie and TV outfit once sponsored what it called the Netflix Prize, asking the world’s data scientists to build new algorithms that could better predict what movies and shows you want to see. And though this certainly advanced the state of the art, Netflix is now exploring yet another leap forward. In an effort to further hone its recommendation engine, the company is delving into “deep learning,” a branch of artificial intelligence that seeks to solve particularly hard problems using computer systems that mimic the structure and behavior of the human brain. The company details these efforts in a recent blog post.

eep learning remained on the outskirts of academia until the mid-aughts, when computers were finally powerful and affordable enough to begin doing some practical work. Although today’s largest neural networks mimic only about one percent of the human brain, according to Engineering and Technology Magazine, modern deep learning algorithms can detect faces in photos, learn your tastes and habits, and, to a certain extent, understand what you say.

Ref: Netflix Is Building an Artificial Brain Using Amazon’s Cloud – Wired