I’M SORRY (information is out-of-date… I’m now updating it slowly).

Introducing: Blocks and Fuel – Frameworks for Deep Learning in Python

Physicists have discovered a material that superconducts at a temperature significantly warmer than the coldest ever measured on the earth. That should herald a new era of superconductivity research.

The world of superconductivity is in uproar. Last year, Mikhail Eremets and a couple of pals from the Max Planck Institute for Chemistry in Mainz, Germany, made the extraordinary claim that they had seen hydrogen sulphide superconducting at -70 °C. That’s some 20 degrees hotter than any other material—a huge increase over the current record.

Last December a paper was posted to arXiv, physicists were cautious about the work. The history of superconductivity is littered with dubious claims of high-temperature activity that later turn out to be impossible to reproduce. But in the months since then, Eremets and co have worked hard to conjure up the final pieces of conclusive evidence. A few weeks ago, their paper was finally published in the peer reviewed journal *Nature*, giving it the rubber stamp of respectability that mainstream physics requires. Suddenly, superconductivity is back in the headlines.

Computers have never been good at answering the type of verbal reasoning questions found in IQ tests. Now a deep learning machine unveiled in China is changing that.

Just over 100 years ago, the German psychologist William Stern introduced the intelligence quotient test as a way of evaluating human intelligence. Since then, IQ tests have become a standard feature of modern life and are used to determine children’s suitability for schools and adults’ ability to perform jobs.

These tests usually contain three categories of questions: logic questions such as patterns in sequences of images, mathematical questions such as finding patterns in sequences of numbers and verbal reasoning questions, which are based around analogies, classifications, as well as synonyms and antonyms.

It is this last category that has interested Huazheng Wang and pals at the University of Science and Technology of China and Bin Gao and buddies at Microsoft Research in Beijing. Computers have never been good at these. Pose a verbal reasoning question to a natural language processing machine and its performance will be poor, much worse than the average human ability.Today, that changes thanks to Huazheng and pals who have built a deep learning machine that outperforms the average human ability to answer verbal reasoning questions for the first time.

Human performance on these tests tends to correlate with educational background. So people with a high school education tend to do least well, while those with a bachelor’s degree do better and those with a doctorate perform best. “Our model can reach the intelligence level between the people with the bachelor degrees and those with the master degrees,” say Huazheng and co.

That’s fascinating work that reveals the potential of deep learning techniques. Huazheng and co are clearly optimistic about future developments. “With appropriate uses of the deep learning technologies, we could be a further step closer to the true human intelligence.”

Deep learning techniques are currently sweeping through computer science like wildfire and the revolution they are creating is still in its early stages. There’s no telling where this revolution will take us but one thing is for sure: William Stern would be amazed.

Ref: arxiv.org/abs/1505.07909 : Solving Verbal Comprehension Questions in IQ Test by Knowledge-Powered Word Embedding

Professor Gui-Qiang G. Chen presents in his inaugural lecture several examples to illustrate the origins, developments, and roles of partial differential equations in our changing world.

While calculus is a mathematical theory concerned with change, differential equations are the mathematician’s foremost aid for describing change. In the simplest case, a process depends on one variable alone, for example time. More complex phenomena depend on several variables – perhaps time and, in addition, one, two or three space variables. Such processes require the use of partial differential equations. The behaviour of every material object in nature, with timescales ranging from picoseconds to millennia and length scales ranging from sub-atomic to astronomical, can be modeled by nonlinear partial differential equations or by equations with similar features. The roles of partial differential equations within mathematics and in the other sciences become increasingly significant. The mathematical theory of partial differential equations has a long history. In the recent decades, the subject has experienced a vigorous growth, and research is marching on at a brisk pace.

University of Oxford Podcasts – Audio and Video Lectures

Downloads “Fully Nonlinear Elliptic Equations book: Fully Nonlinear Elliptic Equations” book by Luis A. Caffarelli bit.ly/zNEhpU