Employees at Spotify threatened to strike if they were not given editorial control over the guests invited on Joe Rogan’s popular podcast. Abigail Shrier’s book Irreversible Damage: The Transgender Craze Seducing Our Daughters was briefly removed from Target’s inventory following complaints, only to be quickly replaced following a backlash. And students at the University of Chicago have sought to have Professor Dorian Abbot publicly censured for criticizing the university’s diversity and inclusion initiatives and the censorious climate on campus. Numerous other examples abound in academia, publishing, news media, and elsewhere. Anxiety about this trend peaked over this summer and led a number of commentators to sign the Harper’s Letter in an effort to defend free speech and inquiry (an initiative that produced its own furious row).
These emotional attempts to suppress controversial or unpopular speech have increasingly made use of what I call the “Mourner’s Veto”—individuals will say that a speaker or a piece of writing has caused them to become distressed or sad or angry or frightened, and they will support these claims with allegations of “harm” or even threats to their “right to exist.” Reasonable debate and discussion then becomes impossible as activists make unfalsifiable but furiously emotive claims about alleged threats to their safety and wellbeing amid much weeping and claims of exhaustion and mental fragility. It is not healthy for the limits of permissible speech to be dictated by the most sensitive person in the room, nor to allow emotional appeals to supplant robust argument as the most effective strategy in a debate.
CEO Tim Cook says that Apple has “things going on in our labs that are mind-blowing” when it comes to fitness. The comment was part of an extended in-person interview at Apple Park in Cupertino, California, that’s featured in the latest episode of the Outside Podcast. Cook and podcast host Michael Roberts spent more than an hour walking the grounds of the company’s headquarters, discussing everything from Cook’s need for daily exercise, to the new breed of health studies enabled by Apple Watch users, to the challenges of escaping screens in today’s hyper-connected world.
In September, Apple introduced the Watch Series 6, with the slogan “The future of health is on your wrist.” Cook told Roberts that the device, which now has a blood-oxygen sensor as well as an electrocardiogram app, is allowing people to “own their health in a way that they were not able to do before.” Later this month, Apple will launch Fitness+, inserting itself into the rapidly expanding online training space in a major way. Fitness+ will integrate your Watch data into guided workouts streamed to an iPhone, iPad, or Apple TV screen. Cook believes that the simple interface will encourage people to explore new kinds of exercise routines and “maybe expand their universe a bit.” This kind of coaching isn’t new for Apple, he argued, but a “broadening” of the relationship the company already has with customers through its retail stores. [audio, 1:10:05]
Duke University researcher Annchen Knodt’s lab published the latest paper challenging the reliability of common brain scan projects, based on about 60 studies of the past decade including her own.
“We found this poor result across the board,” Knodt said. “We’re basically discrediting much of the work we’ve done.”
bioRxiv; Bruce J. Wittmann, Yisong Yue, Frances H. Arnold
from
Due to screening limitations, in directed evolution (DE) of proteins it is rarely feasible to fully evaluate combinatorial mutant libraries made by mutagenesis at multiple sites. Instead, DE often involves a single-step greedy optimization in which the mutation in the highest-fitness variant identified in each round of single-site mutagenesis is fixed. However, because the effects of a mutation can depend on the presence or absence of other mutations, the efficiency and effectiveness of a single-step greedy walk is influenced by both the starting variant and the order in which beneficial mutations are identified—the process is path-dependent. We recently demonstrated a path-independent machine learning-assisted approach to directed evolution (MLDE) that allows in silico screening of full combinatorial libraries made by simultaneous saturation mutagenesis, thus explicitly capturing the effects of cooperative mutations and bypassing the path-dependence that can limit greedy optimization. Here, we thoroughly investigate and optimize an MLDE workflow by testing a number of design considerations of the MLDE pipeline. Specifically, we (1) test the effects of different encoding strategies on MLDE efficiency, (2) integrate new models and a training procedure more amenable to protein engineering tasks, and (3) incorporate training set design strategies to avoid information-poor low-fitness protein variants (“holes”) in the training data. When applied to an epistatic, hole-filled, four-site combinatorial fitness landscape of protein G domain B1 (GB1), the resulting focused training MLDE (ftMLDE) protocol achieved the global fitness maximum up to 92% of the time at a total screening burden of 470 variants. In contrast, minimal-screening-burden single-step greedy optimization over the GB1 fitness landscape reached the global maximum just 1.2% of the time; ftMLDE matching this minimal screening burden (80 total variants) achieved the global optimum up to 9.6% of the time with a 49% higher expected maximum fitness achieved. To facilitate further development of MLDE, we present the MLDE software package (https://github.com/fhalab/MLDE), which is designed for use by protein engineers without computational or machine learning expertise. [full text]
The University of Michigan’s Board of Regents officially has hired an outside firm to help guide its response to sexual assault complaints and help overhaul a culture many said led to the improper handling of accusations against high-profile employees.
Guidepost Solutions is an investigations, regulatory compliance, monitoring and security consulting firm that has served as the independent safety monitor for General Motors and as federal monitor of the New York City Housing Authority. It also has addressed sexual misconduct or discrimination matters at a number of public and private universities.
A former federal prosecutor, Asha Muldro, will lead the team working with U-M.
“The key to our success here at the University of Michigan will be a solid and productive working relationship with the university community. Our focus will be to understand institutional needs, challenges and cultures to define solutions that are impactful and sustainable over the long term,” Muldro told regents Thursday. “This is a forward-looking constructive effort to work with the university in a changing environment.”
As I look out my condo window this early Tuesday morning, I see the city garbage truck enter our village, then stop at the curb. A swing arm emerges, drifts toward a dumpster, hoists it mid-air, and tips it into the dump bed. The robotic arm reverses the sequence, and the truck drives off. My Florida city isn’t alone: many communities have now adopted robotics, often citing the entwined goals of increasing recycling and reducing the cost of hauling away trash. By using automated trucks, the companies collect trash faster with fewer workers, reducing their payroll costs and allowing them to charge communities significantly less for their service. But what happened to those unskilled workers who used to hang off the backs of the trucks and retrieve the trash cans by hand?
Concern about unskilled workers in the age of artificial intelligence (AI), especially in the developing world, is the focus of a recent International Monetary Fund (IMF) working paper. New technologies like AI, machine learning, robotics, big data, and networks are expected to revolutionize production processes, but they could also have a major impact on developing economies. In expectation of these new technologically-driven pressures, a drastic shift to rapidly improve productivity gains and invest in education and skills development is needed to capitalize on the much-anticipated demographic transition associated with AI.
University of Southern California, Viterbi School of Engineering
from
Movie ratings can determine a movie’s appeal to consumers and the size of its potential audience. Thus, they have an impact on a film’s bottom line. Typically, humans do the tedious task of manually rating a movie based on viewing the movie and making decisions on the presence of violence, drug abuse and sexual content.
Now, researchers at the USC Viterbi School of Engineering, armed with artificial intelligence tools, can rate a movie’s content in a matter of seconds, based on the movie script and before a single scene is shot. Such an approach could allow movie executives the ability to design a movie rating in advance and as desired, by making the appropriate edits on a script and before the shooting of a single scene. Beyond the potential financial impact, such instantaneous feedback would allow storytellers and decision-makers to reflect on the content they are creating for the public and the impact such content might have on viewers.
Using artificial intelligence applied to scripts, Shrikanth Narayanan, University Professor and Niki & C. L. Max Nikias Chair in Engineering, and a team of researchers from the Signal Analysis and Interpretation Lab (SAIL) at USC Viterbi, have demonstrated that linguistic cues can effectively signal behaviors on violent acts, drug abuse and sexual content (actions that are often the basis for a film’s ratings) about to be taken by a film’s characters.
Researchers at Oregon State University are making key advances with a new type of optical sensor that more closely mimics the human eye’s ability to perceive changes in its visual field.
The sensor is a major breakthrough for fields such as image recognition, robotics and artificial intelligence. Findings by OSU College of Engineering researcher John Labram and graduate student Cinthya Trujillo Herrera were published today in Applied Physics Letters.
Previous attempts to build a human-eye type of device, called a retinomorphic sensor, have relied on software or complex hardware, said Labram, assistant professor of electrical engineering and computer science. But the new sensor’s operation is part of its fundamental design, using ultrathin layers of perovskite semiconductors – widely studied in recent years for their solar energy potential – that change from strong electrical insulators to strong conductors when placed in light.
“You can think of it as a single pixel doing something that would currently require a microprocessor,” said Labram, who is leading the research effort with support from the National Science Foundation.
It likely would come as no surprise to any teacher of young children that we found that our networks became unstable after continuous periods of learning. However, when we exposed the networks to states that are analogous to the waves that living brains experience during sleep, stability was restored. It was as though we were giving the neural networks the equivalent of a good, long nap.
This sort of instability is not a characteristic of all AI networks. The issue only arises when training biologically realistic processors, or when trying to understand biology itself. The vast majority of researchers on machine learning, deep learning and AI never encounter this instability because, in the very artificial systems they study, they have the luxury of performing mathematical operations that have no equivalent in living neurons.
Our decision to expose our biologically realistic networks to an artificial analogue of sleep was nearly a last-ditch effort to stabilize them. They were spontaneously generating images that were analogous to hallucinations.
The dramatic arrest in 2018 of Joseph James DeAngelo Jr. was all the more astounding because of how detectives said they caught the elusive Golden State Killer — by harnessing genetic technology already in use by millions of consumers to trace their family trees.
But the DNA-matching effort that caught one of America’s most notorious serial killers was more extensive than previously disclosed and involved covert searches of private DNA housed by two for-profit companies despite privacy policies, according to interviews and court discovery records accessed by The Times.
The revelations are likely to heighten debate about genetic privacy and the self-policing models of testing companies, as well as law enforcement access.
Estimating the number of fish in the sea is a wet, cold, and inexact business. To gauge how populations are faring—a critical part of managing fisheries—researchers typically drag a large net behind a ship, counting and measuring what they catch. But these trawl surveys only provide a rough indication of fish populations and they cost tens of thousands of dollars a day. Many researchers are hoping sampling loose bits of DNA that float in seawater can improve the surveys and extend them into places where trawls can’t go: sensitive habitats like coral reefs, wind farms, and stretches of rocky sea floor that are treacherous for heavy nets.
A large study published today boosts confidence that environmental DNA (eDNA) could become a reliable indicator of the abundance of fish. “There is more and more evidence,” says Einar Nielsen, a geneticist at the Technical University of Denmark, who was not involved in the work. But additional research is needed before the technique can be put into practice, he and other experts say.
Carnegie Mellon University, ML@CMU, Machine Learning Blog
from
The International Conference on Machine Learning (ICML) is a flagship machine learning conference that in 2020 received 4,990 submissions and managed a pool of 3,931 reviewers and area chairs. Given that the stakes in the review process are high — the careers of researchers are often significantly affected by the publications in top venues — we decided to scrutinize several components of the peer-review process in a series of experiments. Specifically, in conjunction with the ICML 2020 conference, we performed three experiments that target: resubmission policies, management of reviewer discussions, and reviewer recruiting. In this post, we summarize the results of these studies.
Yale University has launched a campus-wide initiative that will unite institutional leadership and academic experts across the natural sciences, engineering, social sciences, professional schools, and the humanities in an intensive effort to tackle the environmental challenges threatening life on Earth.
The Planetary Solutions Project will identify and advance solutions for an array of the most pressing environmental problems caused by human activities — especially climate change and biodiversity loss — through enhanced resources, expanded collaborations, and boundary-pushing problem-solving.
The University of Maine at Presque Isle has launched an undergraduate degree program in computer science aimed at giving students a running start in that fast-growing professional field.
The school set up the new bachelor of science program via a federal grant received last fall from the U.S. Department of Education Title III Strengthening Institutions Program.
Our daily interactions with technology create vast amounts of data and analytics giving rise to what has been dubbed the “artificial intelligence revolution.” Now, a new research centre at the University of Toronto’s Temerty Faculty of Medicine aims to harness the incredible promise of AI in the realms of medicine and health care.
The Temerty Centre for Artificial Intelligence Research and Education in Medicine (T-CAIREM) launched this week at U of T, solidifying Toronto’s place at the nexus of AI, data science and the health sciences.
“Toronto is uniquely positioned to lead globally in artificial intelligence in healthcare,” says Professor Muhammad Mamdani, who was recently appointed the inaugural director of T-CAIREM for a five-year term. “Our expertise in medicine and allied health sciences, computer science, statistics, mathematics and engineering is among the best in the world.”
Online December 14. “This is the public portion of the Metadata Day 2020 workshop, which convenes leaders from several open source projects for graph-based dataset metadata management, along with thought leaders in metadata management and dataset governance from Google, IBM, UC Berkeley, etc.” [registration required]
“Summer 2020 was an unprecedented year for all. The ongoing pandemic prevented us from welcoming our students to Pittsburgh for our usual in-person program. However, we quickly made a new plan and virtually welcomed nearly 40 students for our first ever remote program! We may not know what the future holds in regards to whether the program will be held in-person or remotely in summer 2021, but we do know that we will provide you with a high quality summer research experience.” Deadline for applications is January 15, 2021.
SPONSORED CONTENT
The eScience Institute’s Data Science for Social Good program is now accepting applications for student fellows and project leads for the 2021 summer session. Fellows will work with academic researchers, data scientists and public stakeholder groups on data-intensive research projects that will leverage data science approaches to address societal challenges in areas such as public policy, environmental impacts and more. Student applications due 2/15 – learn more and apply here. DSSG is also soliciting project proposals from academic researchers, public agencies, nonprofit entities and industry who are looking for an opportunity to work closely with data science professionals and students on focused, collaborative projects to make better use of their data. Proposal submissions are due 2/22.
Google’s UK-based lab and research company DeepMind has added Jraph to the growing number of open-sourced libraries around JAX, while surveying the machine learning framework’s development and ecosystem.
JAX is a Python library that Google researchers developed and introduced in 2018 for high-performance numerical computing. JAX combines NumPy, automatic differentiation, and GPU/TPU support. In a new blog post, DeepMind researchers look at how JAX and its emergent ecosystem of open source libraries have served and accelerated an increasing number of machine learning projects.
