In a spotlight paper from the 2017 NIPS Conference, my team and I presented an AI optimization framework we call Net-Trim, which is a layer-wise convex scheme to prune a pre-trained deep neural network. Deep learning has become a method of choice for many AI applications, ranging from image recognition to language translation. Thanks to algorithmic and computational advances, we are now able to train bigger and deeper neural networks resulting in increased AI accuracy. However, because of increased power consumption and memory usage, it is impractical to deploy such models on embedded devices with limited hardware resources and power constraints. One practical way to overcome this challenge is to reduce the model complexity without sacrificing accuracy. The solution involves removing potentially redundant weights to make the network sparser.

Chattermill, a London-based startup that uses'deep learning' to help companies make better sense of customer feedback, has raised £600,000 in seed funding. Backing comes from Entrepreneur First -- Chattermill is an alumnus of the company builder -- and Avonmore Developments, along with a number of angel investors, including Jeff Kelisky, CEO of Seedrs. Founded in 2015 by friends Mikhail Dubov and Dmitry Isupov, Chattermill is one of a number of startups that are tackling the problem of how to sift through and respond to customer feedback and across multiple channels. With that data growing exponentially, the company is employing deep learning to help do the job in, arguably, a much more scalable and potentially more accurate way. "We help companies understand and improve their customer experience: we give companies insight that helps them craft better products and services," Dubov, Chattermill's CEO, tells me. "Companies with best in class customer experience ultimately have more loyal customers and find it easier acquiring them in the first place.

"Deep Learning" computer systems, based on artificial neural networks that mimic the way the brain learns from an accumulation of examples, have become a hot topic in computer science. In addition to enabling technologies such as face- and voice-recognition software, these systems could scour vast amounts of medical data to find patterns that could be useful diagnostically, or scan chemical formulas for possible new pharmaceuticals. But the computations these systems must carry out are highly complex and demanding, even for the most powerful computers. Now, a team of researchers at MIT and elsewhere has developed a new approach to such computations, using light instead of electricity, which they say could vastly improve the speed and efficiency of certain deep learning computations. Their results appear today in the journal Nature Photonics in a paper by MIT postdoc Yichen Shen, graduate student Nicholas Harris, professors Marin Soljacic and Dirk Englund, and eight others.

Pieter started his invited talk by summarizing some of the key differences between supervised learning and Reinforcement Learning (RL). In essence, RL is mainly concerned with learning an effective policy to have an agent interact with the world in a way that best achieves a goal. For example, learning a policy on how to walk. Recently, RL has seen many success stories, such as learning to play Atari games from the raw pixel inputs, mastering the game of Go to a superhuman level, or effectively teaching simulated characters how to walk from scratch. However, one big gap between RL algorithms and humans, remains the time it takes to acquire new and effective policies.

We are back with some highlights from the second day of NIPS. A lot of fascinating research was showcased today, and we are excited to share some of our favorites with you. If you missed them, feel free to check our Day 1 and Day 3 Highlights! One of the most memorable sessions of the first two days was today's invited talk by Kate Crawford, about bias in Machine Learning. We recommend taking a look at the feature image of this post, representing modern Machine Learning datasets as an attempt at creating a taxonomy of the world.

This talk gave a solid overview of the current state and recent advances in Deep Learning. Convolutional Neural Networks (CNN) and autoregressive models are starting to see ubiquitous use in production, showing a fast transition from research to industry. These models have taught us that introducing inductive biases such as translation invariance (CNN) or time recurrence (Recurrent Neural Networks) can be extremely useful. We've also found out that simple "tricks" such as Residual Networks or Attention can lead to tremendous leaps in performance. There are good reasons to believe we will find more such "tricks".

The summary can be created by calling the summary() function on the model that returns a string that in turn can be printed. Below is the updated example that prints a summary of the created model. Running this example prints the following table. We can clearly see the output shape and number of weights in each layer. The summary is useful for simple models, but can be confusing for models that have multiple inputs or outputs. Keras also provides a function to create a plot of the network neural network graph that can make more complex models easier to understand. The plot_model() function in Keras will create a plot of your network.

In recent years, the amalgam of deep learning fundamentals with Natural Language Processing techniques has shown a great improvement in the information mining tasks on unstructured text data. The models are now able to recognize natural language and speech comparable to human levels. Despite such improvements, discrepancies in the results still exist as sometimes the information is coded very deep in the syntaxes and syntactic structures of the corpus. User: Hi, I took a horrible picture in a museum, can you tell where is it located? User: Hi, I took a horrible picture in a museum, can you tell where is it located?

Athelas (https://athelas.com) is a novel, at-home diagnostics platform that enables pharmaceutical companies to better manage therapies.

Many years ago I was with Garry Kasparov for an event in London's Home House, and there we had dinner with a young lad, a former child prodigy in chess, one who had reached master level (Elo 2300) at the age of 13 and captained a number English junior chess teams. It was an interesting encounter with the boy enthusiastically describing a computer game he was developing. After he left I said to Garry: "That's a cocky young fellow!" "But very smart," Garry replied. And we left it at that. More than twenty years later I had occasion to contact him again.