Deepgram, a startup applying machine learning to audio data, is releasing its machine transcription platform this morning for free. No more will you have to pay for other services like Trint to get the dirty work of automated transcription done. Hint: it has something to do with data. In fact, machine anything isn't solved. And it seems like everyone these days is making haste to build their own Fort Knox of data to solve machine everything.

Google has built a pair of headphones that can translate foreign languages in real time. The Pixel Buds are like a real-world equivalent of the Babel fish, the famous fictional creature from The Hitchhiker's Guide to the Galaxy. They both translate and enable you to speak in foreign languages, and worked incredibly impressively in a demo at Google's launch event this week, enabling an English speaker to hold a smooth conversation with a Swedish speaker. "It's like you've got your own personal translator with you everywhere you go," says Google. "Say you're in Little Italy, and you want to order your pasta like a pro. All you have to do is hold down on the right earbud and say, 'Help me speak Italian'. As you talk, your Pixel phone's speaker will play the translation in Italian out loud. When the waiter responds in Italian, you'll hear the translation through your Pixel Buds."

Since its latest edition, the Google Translate application now supports over 100 languages and serves a worldwide community of over 500 million users virtually. But over 6,000 languages are actually spoken, with around 360 languages spoken by a million people or more. To combat this disparity in translation technology, the USC Information Sciences Institute is working on a universal language translation system. Supported by the Defense Advanced Research Projects Agency, the program involves ISI Director of Natural Language Technologies Kevin Knight and a small team of full-time researchers and doctoral students. Knight has worked with ISI in natural language processing for 25 years, and the system is one of many projects he's working on in the field.

We are in a great time with regards to AI and Machine Learning, due to immense interest and the pace of technological advances. Recent papers show that gradient descent will asymptotically avoid saddle points and it can take exponential time to escape saddle points. Jia Li, Head of R&D, Cloud AI and Machine Learning, Google gave an inspirational keynote on "Why democratizing AI matters: Computing, data, algorithms, and talent". Talking about algorithms, she mentioned that her team switched from phrase-based machine translation to Neural Machine Translation (NMT), and this has led to many improvements.

Last week, experts from the AI world came together for the Artificial Intelligence Conference at San Francisco to discuss insights, opportunities, challenges and trends related to the rapidly expanding field of AI. The conference included hands-on trainings, tutorials, startup showcase (which was won by PipelineAI), keynotes, sessions, expo, and social events. Here is my report on Key Takeaways from AI Conference in San Francisco 2017 – Day 1. Michael Jordan, Distinguished Professor, UC Berkeley gave his keynote on "How to escape saddle points efficiently". We are in a great time with regards to AI and Machine Learning, due to immense interest and the pace of technological advances. However, the theories and our understanding is lagging to keep up with the challenges.

Recent advances in artificial intelligence are encouraging governments and corporations to deploy AI in high-stakes settings including driving cars autonomously, managing the power grid, trading on stock exchanges, and controlling autonomous weapons systems. Such applications require AI methods to be robust to both the known unknowns (those uncertain aspects of the world about which the computer can reason explicitly) and the unknown unknowns (those aspects of the world that are not captured by the system’s models). This article discusses recent progress in AI and then describes eight ideas related to robustness that are being pursued within the AI research community. While these ideas are a start, we need to devote more attention to the challenges of dealing with the known and unknown unknowns. These issues are fascinating, because they touch on the fundamental question of how finite systems can survive and thrive in a complex and dangerous world

Zuckerberg, who spent the summer sparring with Tesla's Elon Musk over the risks of ever-advancing artificial intelligence in our technology, got some support from Google's head of search and AI, John Giannandrea, who spoke recently about what he called the "huge amount of unwarranted hype around AI right now." Speaking at the TechCrunch Disrupt conference in San Francisco on Tuesday, Giannandrea echoed some of the Facebook co-founder's recent statements dismissing doomsday scenarios in which AI-empowered machines pose an inherent existential threat to their human creators. "This leap into, 'Somebody is going to produce a superhuman intelligence, and then there's going to be all these ethical issues' is unwarranted and borderline irresponsible," Giannandrea said at the conference. Google's AI chief added: "I'm definitely not worried about the AI apocalypse." Giannandrea went on to explain the importance of machine learning and artificial intelligence in revolutionizing the technology industry. Google uses AI to power features like Google Translate, the online tool that can instantly translate both spoken words and typed text, as well as products that help users search for new jobs online and provide you with ready-made replies to messages in Google's Gmail, among countless other applications.

We propose a new generative model of sentences that first samples a prototype sentence from the training corpus and then edits it into a new sentence. Compared to traditional models that generate from scratch either left-to-right or by first sampling a latent sentence vector, our prototype-then-edit model improves perplexity on language modeling and generates higher quality outputs according to human evaluation. Furthermore, the model gives rise to a latent edit vector that captures interpretable semantics such as sentence similarity and sentence-level analogies.

Today I want to show an example of Sequence to Sequence model with all the latest TensorFlow APIs [as of TF 1.3]. Seq2Seq models are very useful when both your input and output have some structure or time component. Most popular applications are all in the language domain, but one can use it to process time series, trees, and many other intrinsically structured data. Translation has been domain where this models advanced the most, as it has a large enough dataset to train large and complicated models and provides a clear value from advancing state-of-the-art. If you haven't seen, here are few papers on Neural Language Translation with Seq2Seqs: https://arxiv.org/abs/1409.3215,

We present an approach to automate the process of discovering optimization methods, with a focus on deep learning architectures. We train a Recurrent Neural Network controller to generate a string in a domain specific language that describes a mathematical update equation based on a list of primitive functions, such as the gradient, running average of the gradient, etc. The controller is trained with Reinforcement Learning to maximize the performance of a model after a few epochs. On CIFAR-10, our method discovers several update rules that are better than many commonly used optimizers, such as Adam, RM-SProp, or SGD with and without Momentum on a ConvNet model. We introduce two new optimizers, named PowerSign and AddSign, which we show transfer well and improve training on a variety of different tasks and architectures, including ImageNet classification and Google's neural machine translation system.