This guide is patterned after my “Doing well in your courses”, a post I wrote a long time ago on some of the tips/tricks I’ve developed during my undergrad. I’ve received nice comments about that guide, so in the same spirit, now that my PhD has come to an end I wanted to compile a similar retrospective document in hopes that it might be helpful to some. Unlike the undergraduate guide, this one was much more difficult to write because there is significantly more variation in how one can traverse the PhD experience. Therefore, many things are likely contentious and a good fraction will be specific to what I’m familiar with (Computer Science / Machine Learning / Computer Vision research). But disclaimers are boring, lets get to it!

Source: A Survival Guide to a PhD

Earlier this summer I wrote about Mozilla’s efforts to rollout a multi-process architecture, codename Electrolysis, for Firefox. In the months since, Mozilla has completed its initial tests on 1 percent of its user population and the initial numbers are good, according to Asa Dotzler, director of Firefox at Mozilla.

The company is reporting a 400 percent improvement in responsiveness and a 700 percent improvement in responsiveness for loading large web pages.These numbers mean that users are far less likely to see their browser freeze, pause, lag or crash. Dotzler himself used the word “janky” to describe previous versions of the browser.

Over the next week, multi-process will be coming to 10 percent of total Firefox users. For now, users with add-ons will not be getting the new architecture. The staggered rollout is fairly industry standard to avoid shipping bugs. Having two independent groups of users allows Mozilla to benchmark metrics from the new version against unconverted users.

For now, multi-process is limited to a single content process and a single browser process. Later versions will include multiple content processes and sandboxing.

Source: Multi-process Firefox brings 400-700% improvement in responsiveness | TechCrunch

Short answer: European startups fail to recognize that when they expand to a new market they have to adjust themselves to the rules, standards and requirements of that specific market. Are we all alike?

Source: Why European Startups Fail to Scale? | Arkley Seed Fund SA

How to raise a genius: lessons from a 45-year study of super-smart children

On a summer day in 1968, professor Julian Stanley met a brilliant but bored 12-year-old named Joseph Bates. The Baltimore student was so far ahead of his classmates in mathematics that his parents had arranged for him to take a computer-science course at Johns Hopkins University, where Stanley taught. Even that wasn’t enough. Having leapfrogged ahead of the adults in the class, the child kept himself busy by teaching the FORTRAN programming language to graduate students.

Unsure of what to do with Bates, his computer instructor introduced him to Stanley, a researcher well known for his work in psychometrics — the study of cognitive performance. To discover more about the young prodigy’s talent, Stanley gave Bates a battery of tests that included the SAT college-admissions exam, normally taken by university-bound 16- to 18-year-olds in the United States.

Bates’s score was well above the threshold for admission to Johns Hopkins, and prompted Stanley to search for a local high school that would let the child take advanced mathematics and science classes. When that plan failed, Stanley convinced a dean at Johns Hopkins to let Bates, then 13, enrol as an undergraduate.

Source: How to raise a genius: lessons from a 45-year study of super-smart children : Nature News & Comment

To-do lists get a lot of flack, but the simple act of planning has some psychological and productivity benefits all by itself.

For a long time, I resisted to-do lists. I wanted the flexibility. I felt that if I kept a list, it would tie me down to a particular set of tasks. Gradually, though, I came around. The busier my work life became, the more crucial it was to have some sort of running agenda on hand. Before long, I even started adding some of those items onto my weekly calendar. In other words, I’d reluctantly become a planner.

Looking back, it shouldn’t have been so difficult. In fact, there are at least three psychological benefits to the simple act of drawing up a list of top-priority tasks—whether or not you actually accomplish them.

Writing Makes Your Memory’s Job Easier

Keeping a list of tasks you need to perform is like taking notes when you’re reading a book or listening to a lecture. When you take notes, you need to filter external information, summarize it in your head, and then write it down. Many studies have shown that note taking helps us distill the information we hear and remember it better than we would if we’d just heard or read it.

Source: How Writing To-Do Lists Helps Your Brain (Whether Or Not You Finish Them) | Fast Company | Business + Innovation

Modern archiving technology cannot keep up with the growing tsunami of bits. But nature may hold an answer to that problem already.

Source: How DNA could store all the world’s data : Nature News & Comment

The open source Git project has just released Git 2.10.0, with features and bugfixes from over 70 contributors. Here’s our look at some of the most interesting new features:

Progress reporting for pushes

When you run git push, you’ve probably seen a progress meter telling you how many objects you’ve sent, how many are left, and how fast the data is moving. But what happens after all of the data has made it to the server? Are we done?

Not quite. Even though the receiver of a push does as much work as possible while the data is flowing in, there are a few CPU-intensive tasks it can’t start until the whole thing has arrived. And while that’s happening, Git is completely silent. Most pushes are small enough that this phase finishes quickly, and you never notice. But when pushing a large number of objects, this can take many seconds or even minutes, leaving you to wonder if things are still working.

Worse, because the network connection is completely silent during this phase, you run the risk of the connection being dropped by HTTP proxies or other network infrastructure. That’s an easy way to turn your wondering into frustration.

Git 2.10 adds progress reports for these post-receive operations, to keep you entertained and to make sure the network knows we’re still going.

Source: Git 2.10 has been released · GitHub

Alzheimer’s treatment appears to alleviate memory loss in small trial

A drug called aducanumab might remove the toxic proteins thought to trigger Alzheimer’s disease from the brain, suggests findings from a small clinical trial.

The results, reported on 31 August in Nature¹, showed that aducanumab broke up amyloid-β proteins in patients with early-stage Alzheimer’s disease. The trial mainly tested the safety of the drug in people, and so the final word on whether aducanumab works to ameliorate the memory and cognitive losses associated with Alzheimer’s will have to wait until the completion of two larger phase III trials.  They are now in progress, and planned to run until at least 2020.

The latest study involved 165 people split into different groups, some of which received the drug and one which received a placebo. In the group receiving infusions of aducanumab, 103 patients given the drug once a month for up to 54 weeks experienced a reduction in the amount of tangled amyloid-β in their brains. These results echoed the findings of a pretrial mouse study — reported in the same paper¹ — in which the drug seemed to clear amyloid-β plaques from the animals’ brains.

Source: Alzheimer’s treatment appears to alleviate memory loss in small trial : Nature News & Comment

When we first set out to apply deep learning to genomics, we asked ourselves what the current state of the art is. What problems are researchers working on and what approaches are they using? This post contains a summary of what we found — an overview of popular network architectures in genomics, the types of data used to train deep models, and the outcomes predicted or inferred.

Despite being able to sequence the genome at nucleotide-level resolution, and the abundance of publicly available labeled datasets from sources like the 1000-genome project, ENCODE and GEO, we are still far from bridging the genotype-phenotype divide or predicting disease from genome sequences. This talk by Brendan Frey puts the deep learning-and-genomics problem in context, explaining why sequencing more genomes may not be the answer. The genome is complex and contains many interacting information layers. Most current approaches involve developing a system to interpret the genomic code or a part of it, rather than directly training a network that predicts phenotype from sequence. Below are some of the ways that deep learning has been used for genomics, with emphasis on implementations for the human genome or transcriptome.

Source: Neural Networks for Genomics

It’s a chilling prospect, but the AI we’ve created could transform human nature, argues this spellbinding new book by the author of Sapiens

At the heart of this spellbinding book is a simple but chilling idea: human nature will be transformed in the 21st century because intelligence is uncoupling from consciousness. We are not going to build machines any time soon that have feelings like we have feelings: that’s consciousness. Robots won’t be falling in love with each other (which doesn’t mean we are incapable of falling in love with robots). But we have already built machines – vast data-processing networks – that can know our feelings better than we know them ourselves: that’s intelligence. Google – the search engine, not the company – doesn’t have beliefs and desires of its own. It doesn’t care what we search for and it won’t feel hurt by our behaviour. But it can process our behaviour to know what we want before we know it ourselves. That fact has the potential to change what it means to be human.

Yuval Noah Harari’s previous book, the global bestseller Sapiens, laid out the last 75,000 years of human history to remind us that there is nothing special or essential about who we are. We are an accident. Homo sapiens is just one possible way of being human, an evolutionary contingency like every other creature on the planet. That book ended with the thought that the story of homo sapiens could be coming to an end. We are at the height of our power but we may also have reached its limit. Homo Deus makes good on this thought to explain how our unparalleled ability to control the world around us is turning us into something new.

The evidence of our power is everywhere: we have not simply conquered nature but have also begun to defeat humanity’s own worst enemies. War is increasingly obsolete; famine is rare; disease is on the retreat around the world. We have achieved these triumphs by building ever more complex networks that treat human beings as units of information. Evolutionary science teaches us that, in one sense, we are nothing but data-processing machines: we too are algorithms. By manipulating the data we can exercise mastery over our fate. The trouble is that other algorithms – the ones that we have built – can do it far more efficiently than we can. That’s what Harari means by the “uncoupling” of intelligence and consciousness. The project of modernity was built on the idea that individual human beings are the source of meaning as well as power. We are meant to be the ones who decide what happens to us: as voters, as consumers, as lovers. But that’s not true any more. We are what gives networks their power: they use our ideas of meaning to determine what will happen to us.

Source: Homo Deus by Yuval Noah Harari review – how data will destroy human freedom | Books | The Guardian