Data Science newsletter – April 9, 2021

The aim of this systematic review and meta-analysis was to evaluate the characteristics and effectiveness of interventions involving contemporary mobile apps or physical activity trackers (ie, enabling automated and continuous self-monitoring and feedback) in promoting physical activity, as well as in improving engagement and retention, in adults (18–65 years old) without chronic disease. A secondary aim was to explore and compare the effect of specific features in these interventions using metaregression. [full text]

Google research manager Samy Bengio said on Tuesday he was resigning, according to an internal email seen by Reuters, in a blow to the Alphabet Inc unit after the firings of his colleagues who questioned paper review and diversity practices.

Though at least two Google engineers had earlier resigned to protest the dismissal of artificial intelligence (AI) researcher Timnit Gebru, Bengio is the highest-profile employee yet to depart.

The list of universities requiring vaccinations to return to campus in the fall is growing longer by the day.

Why it matters: With the mandates, universities are going where most corporations have not. The political and legal blowback is already taking shape.

The Andrew W. Mellon Foundation awarded $1.4 million to Michigan State University for Enslaved: Peoples of the Historical Slave Trade, or Enslaved.org, a first-of-its-kind database containing millions of records cataloging the lives of enslaved Africans and their descendants.

Today we are pleased to announce that in this first year of the program we have granted 77 awards, which included 86 principal investigators representing 15+ countries and over 50 universities. Of the 86 award recipients, 43% identify as an historically marginalized group within technology. Please see the full list of 2021 recipients on our web page, as well as in the list below.

Humans use different domain languages to represent, explore, and communicate scientific concepts. During the last few hundred years, chemists compiled the language of chemical synthesis inferring a series of “reaction rules” from knowing how atoms rearrange during a chemical transformation, a process called atom-mapping. Atom-mapping is a laborious experimental task and, when tackled with computational methods, requires continuous annotation of chemical reactions and the extension of logically consistent directives. Here, we demonstrate that Transformer Neural Networks learn atom-mapping information between products and reactants without supervision or human labeling. Using the Transformer attention weights, we build a chemically agnostic, attention-guided reaction mapper and extract coherent chemical grammar from unannotated sets of reactions. Our method shows remarkable performance in terms of accuracy and speed, even for strongly imbalanced and chemically complex reactions with nontrivial atom-mapping. It provides the missing link between data-driven and rule-based approaches for numerous chemical reaction tasks.

Working safely is not only about processes, but context – understanding the work environment and circumstances, and being able to predict what other people will do next. A new system empowers robots with this level of context awareness, so they can work side-by-side with humans on assembly lines more efficiently and without unnecessary interruptions.

Instead of being able to only judge distance between itself and its human co-workers, the human-robot collaboration system can identify each worker it works with, as well as the worker’s skeleton model, which is an abstract of the worker’s body volume, says Hongyi Liu , a researcher at KTH Royal Institute of Technology. Using this information, the context-aware robot system can recognize the worker’s pose and even predict the next pose. These abilities provide the robot with a context to be aware of while interacting.

Donatella Danielli became the new director of Arizona State Univesity’s School of Mathematical and Statistical Sciences on Jan. 1. Before coming to ASU, she was a professor of mathematics at Purdue University, where she spent most of her career. She earned her PhD in mathematics from Purdue in 1999 and laurea cum laude from the University of Bologna in 1989.

“We were thrilled to bring Donatella to ASU. She brings a wonderful blend of qualities as a leader for the School of Mathematical and Statistical Sciences,” said Nancy Gonzales, provost pro tempore. “She is a highly accomplished mathematician, a dedicated educator with a lifelong commitment to student success, and she pushes boundaries in pursuit of excellence and inclusion as linked goals. She is founder and co-editor of the newly launched, flagship journal of the Association for Women in Mathematics. She is exactly what we need to take the school to the next level.”

A new pioneer centre for artificial intelligence research is expected to open in Copenhagen at the end of 2021. Headed by world-leading American AI researcher, Serge Belongie, and with a good DKK 350 million to back it

CPU algorithm trains deep neural nets up to 15 times faster than top GPU trainers

Rice University computer scientists have demonstrated artificial intelligence (AI) software that runs on commodity processors and trains deep neural networks 15 times faster than platforms based on graphics processors.

“The cost of training is the actual bottleneck in AI,” said Anshumali Shrivastava, an assistant professor of computer science at Rice’s Brown School of Engineering. “Companies are spending millions of dollars a week just to train and fine-tune their AI workloads.”

Coursera, the online learning platform that allows you to take classes offered by top-ranked schools like Yale and Duke has partnered with Google to introduce the Google Professional Certificates program.

These multi-course curricula are being offered in the fields of Data Analytics, Project Management, UX (user experience) Design, IT Support, and IT Automation. Students who complete any of the programs will receive a certificate that can bolster a resume, or make them eligible for a new career. Google says 82% of U.S. graduates have reported a positive career outcome (new job, promotion, raise) within six months.

This collaboration is part of Google’s “Grow With Google” initiative, which aims help people looking to make a career change or boost their resume.

Atlantic cod was once one of the most important fish species in the world. The big, long-lived predator helped feed Europe and North America for centuries. But decades of overfishing in the mid–20th century caused populations to crash, wiping out 30,000 jobs in Canada alone and financially devastating many coastal communities.

Now, a new study provides a ray of hope for the remaining fish. Researchers have found the cod haven’t lost the genetic diversity that would be crucial for their recovery, something that often happens when species hit a so-called population “bottleneck.” The new research—the first to compare whole genomes of cod from before and after intensive fishing—reinforces the notion that more protective management will help beleaguered stocks bounce back, experts say.

I’m excited to talk about the promises and pitfalls of emotion-sensing A.I. with Luke Stark, assistant professor in the faculty of information and media studies at the University of Western Ontario. For many years, Luke has been studying this topic, sparked widespread discussion by comparing facial recognition technology to plutonium, and has an exciting new book coming out with MIT Press, Ordering Emotion: Histories of Computing and Human Feelings From Cybernetics to AI. In addition to being a renowned scholar who studies ethical issues, Luke prioritizes ethical action. He recently turned down a prestigious and lucrative Google research scholar award as a gesture of solidarity with former Google employees Timnit Gebru and Margret Mitchell.

Dartmouth College is taking major steps to enhance communication with students and faculty, deploying an AI solution that offers a personalized experience in real-time. The College has partnered with artificial intelligence company Aisera to provide students and faculty with a virtual assistant designed to improve communication and response times while learning and teaching at home.

Dartmouth’s virtual assistant named Dart offers support services and provide responses within seconds to approximately 10,000 students and faculty. Dartmouth leaders were looking for a virtual assistant to instantly answer support requests in natural language on a channel of choice. Dartmouth users now interact directly with the virtual assistant within Slack or on their client services portal.

Interviews with campus officials and health administrators around the country reveal similar thinking. Almost every official who spoke with KHN said universities will open their classrooms and their dorms this fall. In many cases, they no longer can afford not to. But controlling those environments and limiting viral spread loom among the largest challenges in many schools’ histories — and the notion of what constitutes normalcy is again being adjusted in real time.

Online April 15, starting at 7 p.m. Eastern. “The event features a panel of three Cornell faculty members and postdoctoral fellows and is moderated by Andrew Ross Sorkin.” [registration required]

Facebook AI has built and open-sourced a new, unique data set called Casual Conversations, consisting of 45,186 videos of participants having nonscripted conversations. It serves as a tool for AI researchers to surface useful signals that may help them evaluate the fairness of their computer vision and audio models across subgroups of age, gender, apparent skin tone, and ambient lighting.

Denison University announces Mathematics and Computer Science faculty Thomas Bressoud and David White have written a new data science textbook, titled “Introduction to Data Systems.” The textbook, which is being incorporated into classrooms at Denison and other universities, is published by international science publisher, Springer Publishing.

… Given the proliferation of smart watches and wristbands, we felt that this was a practical device to investigate to provide heart monitoring. All smart watches have an embedded Bluetooth communication system to receive and transmit data at various ranges. Remote monitoring typically involves collecting information from wearable devices at the edge, integrating that information into a data lake, and generating inferences that can then be served back to the relevant stakeholders. Additionally, in some cases, compute and inference must also be done at the edge to shorten the feedback loop between data collection and response.

You can use Amazon DynamoDB to capture high velocity data IoT device data to develop insights on fitness monitoring. It’s important to store these data points in a centralized data lake in real time, where it can be transformed, analyzed, and combined with diverse organizational datasets to derive meaningful insights and make predictions. As customers are using more and more wearable devices, companies are looking for ways to store high frequency data and perform meaningful analytics.

The Chirpy Cardinal team effectively combined scripted and neurally generated bot responses for richer interactions.