NYU Data Science newsletter – February 16, 2016

GROUP CURATION: N/A

Of all the application areas genomics has affected in precision medicine, clinical diagnostics particularly stands out as having achieved extraordinary progress. More specifically, genetic tests and biomarkers, which are often used in some prognostic manner or for shaping the therapeutic regimen, have been the most scrutinized for their validity and utility to overall patient care.

“Every product needs to prove clinical utility,” declares Harry Glorikian, senior executive, board director, and consultant in the life sciences/healthcare industry. “Does the product/device/test change the clinical management of the patient? A ‘Yes’ is needed to have a product adopted.”

If there was still any doubt about the value of data analytics, this should lay it to rest in a single word: beer.

That’s right, digital marketing agency Havas Helia has created 0101, a beer that was crafted based on data.

We have just completed another round of the Google Research Awards, our annual open call for proposals on computer science and related topics including machine learning, speech recognition, natural language processing, and computational neuroscience. Our grants cover tuition for a graduate student and provide both faculty and students the opportunity to work directly with Google researchers and engineers. This round we received 950 proposals, an increase of 18% over last round, covering 55 countries and over 350 universities. After expert reviews and committee discussions, we decided to fund 151 projects.

Virus study used data released online without adequate acknowledgement, researcher complains.

If the bad news is that we’re living in a world in which resilience is more critical to survival than ever, the good news is that technology is more than ever providing the tools we need to cultivate resilience. Exciting innovations in digital data collection, analysis and visualization now allow us to track and understand human impacts at global to local scales and identify big-picture patterns and processes in ways never before possible, from the National Science Foundation’s Ocean Observatories Initiative, which measures physical, chemical, biological and geological variables throughout the depths of the ocean to the Global Earth Observation System of Systems, which provides petabytes of environmental data from space-borne, airborne and in situ sensors.

Indeed, we now find ourselves inhabiting a “Digital Earth” composed of technologies from satellites to wristwatches that monitor, map, model and manage virtually everything around us.

… Four years ago, organizations and executives were struggling to understand the opportunity and business impact of Big Data. While many executives loathed the term, others were apostles of the belief that data-driven analysis could transform business decision-making. Now, we have arrived at a new juncture: Big Data is emerging as a corporate standard, and the focus is rapidly shifting to the results it produces and the business capabilities it enables. When the internet was a new phenomenon, we’d say “I am going to surf the World Wide Web” – now, we just do it. We are entering that same phase of maturity with Big Data.

So, how can executives prepare to realize value from their Big Data investments?

DataRobot has built a software platform to bring machine learning and data-science tools to businesses. The product is aimed not just at data scientists, but also at software developers, business analysts, and statisticians. The company is led by CEO and co-founder Jeremy Achin, a data scientist and UMass Lowell grad who previously worked at Travelers Insurance.

For nearly half a century, AI has always seemed as if it just beyond reach, less, than two decades away. Yet “strong AI” in some ways still seems elusive. In this talk, I will give a cognitive scientist’s perspective on AI. What have we learned, and what are we still struggling with? Is there anything that researchers of AI can still learn from studying the science of human cognition?

Speaker: Gary Marcus

Thursday, February 18, at 7 p.m., Pivotal Labs (625 Avenue of Americas, 2nd Floor)

Hi Reddit!

We are computer scientists and ethicists who are examining the societal, ethical, and labor market implications of increasing automation due to artificial intelligence. (Bart Selman, Moshe Vardi, Wendell Wallach)

SciPy 2016, the fifteenth annual Scientific Computing with Python conference, will be held this July 11th-17th in Austin, Texas. SciPy is a community dedicated to the advancement of scientific computing through open source Python software for mathematics, science, and engineering. The annual SciPy Conference allows participants from academic, commercial, and governmental organizations to showcase their latest projects, learn from skilled users and developers, and collaborate on code development.

Deadline to submit papers is Monday, April 4.

In a world of search engines and information indexing, some minority languages are waiting their turn. For 12 million Albanian-speakers, it’s also a more fundamental matter of getting documents and books and newspapers and bus schedules that are in Albanian online. That’s where Gjirafa comes in — and the $2 million they just scored in venture capital. … Gjirafa’s founder, CEO Mergim Cahani, started developing his company’s mission when he was a student in the U.S. … While getting his Master’s Degree at New York University, he worked with Professor Torsten Suel, a pioneer in web search that Gjirafa now counts among its top advisers.

Steps to making data reusable range from the organizational (document the contents of your files) to the complex (ensuring the computing environment of your data can be re-created). The first is sometimes more of a matter of convenience: a data file may be usable, but a fellow researcher cannot spare the many hours it might take to sort out how to use poorly documented materials. In other cases, badly documented data can be more than an inconvenience. It can make the entire dataset unusable by you or your colleagues because the context for how the data was collected or how it models your real-world referent cannot be recalled.

Scikit Learn is a new easy-to-use interface for TensorFlow from Google based on the Scikit-learn fit/predict model. Does it succeed in making deep learning more accessible?

Welcome! This ipython notebook (or associated python script GW150914_tutorial.py ) will go through some typical signal processing tasks on strain time-series data associated with the LIGO GW150914 data release from the LIGO Open Science Center (LOSC).