🤖 MLguru #11: Free Machine Learning Webinar, Bypassing Apple FaceID, and The Global PyTorch Summer Hackathon

During the webinar you will learn:

• how to use machine learning in the most effective way,
• how to avoid mistakes,
• how to minimize the risks connected to machine learning solutions.

Of course, once you subscribe, take a look at our ML news from the industry :)

Enjoy the read,

Mateusz

Researchers Bypass Apple FaceID Using Biometrics ‘Achilles Heel’

During Black Hat USA 2019 researchers demonstrated an attack that allowed them to bypass a victim’s FaceID and log into their phone simply by putting a pair of modified glasses on their face. By merely placing tape carefully over the lenses of a pair of glasses and putting them on the victim’s face the researchers demonstrated how they could bypass Apple’s FaceID in a specific scenario. The attack itself is difficult, given that the bad actor would need to figure out how to put the glasses on an unconscious victim without waking them up but, as shown, it’s fully possible. Read more

Win over $60,000 in cash prizes during The Global PyTorch Summer Hackathon

PyTorch is an open source machine learning framework that is used by both researchers and developers to build, train, and deploy ML systems that solve many different complex challenges. The PyTorch team invites you to hack with the PyTorch community to build innovative, impactful models, applications, and other projects that create positive impact for businesses or people. Developers will have the opportunity to win over $60,000 in cash prizes and more. First-place winners will also receive round-trip airfare to attend the PyTorch Developer Conference, and have their projects featured in the event. More info.

Using AI to give doctors a 48-hour head start on life-threatening illness

Millions of people die every year from diseases that could have been prevented with earlier detection. One such disease is acute kidney injury (AKI), a condition where a patient’s kidney suddenly stops working properly. Artificial intelligence can now predict one of the leading causes of AKI up to two days before it happens. Read more.

A new model for word embeddings that are resilient to misspellings

Misspelling Oblivious Embeddings (MOE) is a new model created by Facebook developers. It combines the open source library fastText with a supervised task that embeds misspellings close to their correct variants. Existing word embedding methods are often unable to deal with malformed texts. In real-world tasks, input text generated by people often contains misspellings (a common source of OOV words), particularly in situations like web searches, chat messages, and social media posts. Misspellings appear in up to 15 percent of web search queries. MOE outperforms fastText for user-generated text and is a perfect solution for slang, abbreviations and miswritten phrases. Read more.

Nvidia GPU accelerators get a direct pipe to big data

Nvidia has unveiled GPUDirect Storage, a new capability that enables its GPUs to talk directly with NVM-Express storage. The technology uses GPUDirect’s RDMA facility to transfer data from flash storage into the GPU’s local memory without needing to involve the host CPU and system memory. If successful, Nvidia could edge out CPUs from yet another fast-growing application area. The data science and machine learning market for servers is thought to be a $20 billion to $25 billion per year opportunity, which is about the same size as the combined HPC and deep learning server market. Read more.