NYU Data Science newsletter – September 21, 2016

GROUP CURATION: N/A

Brain-inspired computing seeks to develop new technologies that solve real-world problems while remaining grounded in the physical requirements of energy, speed, and size. Meeting these challenges requires high-performing algorithms that are capable of running on efficient hardware. Here, we adapt deep convolutional neural networks, which are today’s state-of-the-art approach for machine perception in many domains, to perform classification tasks on neuromorphic hardware, which is today’s most efficient platform for running neural networks. Using our approach, we demonstrate near state-of-the-art accuracy on eight datasets, while running at between 1,200 and 2,600 frames/s and using between 25 and 275 mW.

“A new method uses data from satellite and street-level images, such as the ones you can see in Google Maps, to automatically create an inventory of street trees that cities may use to better manage urban forests.”

More Data for Good:

Matchmaking to Find Homes for Refugees (September 19, Bloomberg View, Alex Teytelboym & Scott Duke Kominers)

P2P insurance firm Lemonade launches out of stealth, powered by chatbots, morals, and big bucks (September 21, VentureBeat, Paul Sawers)

Computational Sustainability Virtual Seminar – Measuring progress towards sustainable development goals with machine learning (takes place on September 27, CompSustNet)

Data for Good Exchange 2016 event (takes place on September 25, Bloomberg L.P.)

I list seven possible areas which I believe could be the answer to this question; these answers are highly subjective and biased and they may be all wrong, but hopefully they do contain some interesting pointers for everyone.

The point of this exercise is to show that there are many strands of active research that could result in major AI advances. So here are they, the seven areas where a major general purpose AI breakthrough could come from.

1. Composable Differentiable Architectures (aka Deep Learning)

The following 20 companies are playing a role in shaping the future of artificial intelligence and its capabilities. We featured artificial inteligence companies that are particularly noteworthy as well as those that have invested significantly in artificial intelligence. These AI companies are listed alphabetically.

1. AIBrain

The disappointing results of clinical trials will no longer be able to languish unpublished, thanks to rules released on 16 September by the US Department of Health and Human Services (HHS) and the US National Institutes of Health (NIH).

The solution to these frequent mismatches is a comprehensive, algorithm-based matching system. In the past half-century, matching systems have been used to match medical residents to hospitals, connect kidney donors with dialysis patients, and assign military cadets to branches of service. In our work with the University of Melbourne’s David Delacrétaz and Will Jones from Royal Holloway, we have identified how refugee–community matching systems might improve refugee outcomes by building on the systems many cities worldwide use to assign children to schools.

We decided to conduct a little experiment. On Monday, in partnership with Siena College, the Upshot published a poll of 867 likely Florida voters. Our poll showed Hillary Clinton leading Donald J. Trump by one percentage point.

We decided to share our raw data with four well-respected pollsters and asked them to estimate the result of the poll themselves.

Computer scientists can prove certain programs to be error-free with the same certainty that mathematicians prove theorems. The advances are being used to secure everything from unmanned drones to the internet.

We provide an overview of PSI (“a Private data Sharing Interface”), a system we are developing to enable researchers in the social sciences and other fields to share and explore privacy-sensitive datasets with the strong privacy protections of differential privacy.

Illumina sells sequencers to hospitals and labs and consumer-style tests to individuals. “With spin-off investments Grail and Helix, and a new software-savvy CEO, Illumina is poised to make DNA an even bigger part of your life.” But what is at stake when a private company is this dominant in a market with huge implications for public good?

Non-profit models for large-scale health science:

Allen Institute publishes highest resolution map of the entire human brain to date (September 15, Allen Institute for Artificial Intelligence)

A Diversity of Genomes – New DNA from understudied groups reveals modern genetic variation, ancient population shifts (September 21, Harvard Medical School)

Deadline for abstract submission is Monday, 28 September 2016.

Through the initiative, grants of up to $150,000 will be awarded in support of innovative social science research that brings new data and methods to bear on questions of interest in the foundation’s core programs in Behavioral Economics; the Future of Work; Race, Ethnicity, and Immigration; and Social Inequality. Deadline for Letters of Inquiry is Wednesday, November 30.

Leiden, The Netherlands The workshop will be held on May 8-12, 2017 at the Lorentz Center. Deadline for applications to participate in the Python in Astronomy 2017 workshop is Friday, December 9.

  Upcoming No-Cost Data Science Events at NYU and in NYC

• 
  Accurate de novo design of hyperstable constrained peptides
  
  
  on September 14
  Naturally occurring, pharmacologically active peptides constrained with covalent crosslinks generally have shapes that have evolved to fit precisely into binding pockets on their targets. Such peptides can have excellent pharmaceutical properties, combining the stability and tissue penetration of small-molecule drugs with the specificity of much larger protein therapeutics. The ability to design constrained peptides with precisely specified tertiary structures would enable the design of shape-complementary inhibitors of arbitrary targets. Here we describe the development of computational methods for accurate de novo design of conformationally restricted peptides, and the use of these methods to design 18–47 residue, disulfide-crosslinked peptides, a subset of which are heterochiral and/or N–C backbone-cyclized. Both genetically encodable and non-canonical peptides are exceptionally stable to thermal and chemical denaturation, and 12 experimentally determined X-ray and NMR structures are nearly identical to the computational design models. The computational design methods and stable scaffolds presented here provide the basis for development of a new generation of peptide-based drugs.
  
• 
  Meet the MSDS class of 2018! Part 1: Maya Bidanda
  
  
  on September 19
  … Within the field of data science, there are many different disciplines, and an even wider set of possible applications. Can you talk about the subsets of data science, or the data science applications that you’re most interested in perusing?

  I am particularly interested in taking machine learning courses, as well as statistics classes in the economics department. I have tinkered with machine learning in my research?—?to help better identify trends in capital markets?—?and I want to further this work.
  

• 
  New Student Interviews: Eduardo Fierro Farah
  
  
  on September 20
  … Within the field of data science, there are many different disciplines, and an even wider set of possible applications. Can you talk about the subsets of data science, or the data science applications that you’re most interested in perusing?

  As an economist, and as a political scientist, I am particularly interested in the applications that data science has in these fields. Political campaigns, political speeches, and public opinion have always interested me, and I want to use natural language processing as a way of analyzing public policies. Data science also has some really interesting applications in the field of micro-economics, particularly when studying inequality and poverty.

  But I also have a strong passion for sports, and would be excited about the possibility of using data science to analyze soccer, football, or formula one racing.
  

“I describe random forests…and then do a demo using the R randomForest library.”

In this post, I go over five methods to visualize gaps in your line data with D3 and analyze the pros and cons of each. This exploration led to my creation of a D3 plugin called d3-line-chunked, which allows you to easily visualize gaps in your data and has good animation support.

Our work is presented here as a set of software citation principles, a discussion of the motivations for developing the principles, reviews of existing community practice, and a discussion of the requirements these principles would place upon different stakeholders. Working examples and possible technical solutions for how these principles can be implemented will be discussed in a separate paper.

How to cite R and R packages in publications.

“We describe a suite of techniques for expressing causal inference tasks from observational data in SQL….we introduce several optimization techniques that significantly” increase speed.