suchow

It’s rare for a large fraction of grant money to be siphoned to a Dean’s discretionary fund. Typically some smallish fraction (maybe 10%, often significantly less) of indirect costs (which, depending on the funder and the negotiated F&A rate may be anywhere from 0% or nearly 100% on top of the direct costs that fund the research staff and materials, etc.) goes back to the subdivision overseen by that Dean to do with as they please. Everywhere I’ve worked, that amounts to a few (low single digit) percent of total costs being used in the way you describe. And many places return none of indirects to the unit overseeing the PI and so then it’s a cool zero percent.

Do you know what would motivate this behavior? Naively, any rent is more than zero rent, so why let a unit sit empty?

The similarity you see between this DeepMind project, the DARPA program, and research in Tenenbaum's lab is not incidental: there's a steady stream of crosstalk and cross-training between machine learning researchers who engineering artificial intelligences and cognitive scientists who reverse-engineer human intelligences. (Note, for example, that Peter Battaglia, one of the co-authors of this DeepMind project, was a postdoc with Tenenbaum.)

At many universities (in the U.S. at least), on paper, it’s the technology commercialization department that makes the call about pursuing patent protection and the burden is on the PI to report all potentially patentable inventions before release so that the university has time to make the determination. In the case you describe, they’d in principle find you in violation of that policy, but in practice can’t because they don’t know about it unless you tell them.

Can you share a sample email? It’s hard to comment without seeing at least the gist of what’s being written.

Also known as automatic differentiation or autodiff.

You say each letter separately; it’s an acronym and stands for Proceedings of the National Academy of Sciences.

Here’s a good resource https://www.nytimes.com/interactive/2014/upshot/buy-rent-cal..., but a financial advisor who knows your situation and is familiar with local real estate may be better.

This is an excellent example of why inferring causation from an observed correlation requires great care. Note that the study did not randomly assign valedictions to emails and observe the causal impact on response rate. Rather, they observed a correlation between the sign-off chosen and the response rate.

"Thanks" garners many responses because people use it in emails that make reasonable requests with a good chance of response.

My favorite valediction is still "I am, &c.," which is short for "I am your humble and obedient servant". Not sure whether my colleagues appreciate it as much as I do...

The parent comment says that, of all the information-finding tools that hardly anyone uses, library resources are the best, not that library resources are the best information-finding tools; there's no contradiction between what the two of you have noticed.

To boost profits you need to extract more value, regardless of whether you provide more.

The word "smarter" masks some of the complexity here. If you define it as "higher performance on a task widely considered to require intelligence", then we've had computers that are smarter than humans for at least decades. If you define it as "higher performance on every task widely considered to require intelligence", then I'll take that bet, please.

The criteria for winning the prize depend more on the outcome of the research (importance) than its process (originality):

"The said interest shall be divided into five equal parts, which shall be apportioned as follows: /- – -/ one part to the person who shall have made the most important discovery within the domain of physiology or medicine …” (Excerpt from the will of Alfred Nobel)".

See also “The virtues of moderation”. https://digitalcommons.law.yale.edu/yjolt/vol17/iss1/2/

I've had good experiences with PQ (https://pypi.org/project/pq/). Any event that generates a notification triggers adds an entry to the queue. Worker processes get entries from the queue. The queue is stored as another table in your database whose structure and content is managed by PQ, though you can always read/write to it if you want. PQ handles the concurrency.

FYI, Battery Park City is part of Manhattan and was designed in the 1960s to 1980s, with the land created during excavation for the World Trade Center.

Is there a usage error in the first sentence or has English lost the blog / blog post distinction?

Whether peer review is blinded depends on the field. Some of the major ML conferences like NIPS are blinded, but in general most journals are not. The reviewers see the authors and affiliations and the editors see everything.

The misdirection is being used as a rhetorical device — you're supposed to feel a brief confusion when you get to the colon; it's then quickly resolved.

Calling a group of women "you guys" is different from calling a person of unknown gender "guy" because in the former case the gender-neutral usage is the only possible interpretation, whereas in the latter it might either be gender-neutral usage or an assumption of the commenter's gender. Given that this ambiguity is nearly always a distraction from the primary purpose of the conversation (as it is here), rightfully or not, pragmatically it is better to use terms that steer clear of it. See Bryan Garner's Modern English Usage for fuller advice on safely navigating this linguistic and sociocultural terrain.

By asking this question, I assume you are being sarcastic and indirectly suggesting that legislation or other means of enforcing AI safety is impossible because it would refer to matrix multiplication and gradient descent and therefore be unreasonably broad, ruling out many harmless computations. However, it's unlikely that legislation or other enforcement would operate at that level of description, in the same way that laws regarding murder do not reference patterns of motor-neuron activation. It is reasonable to prevent certain certain classes of multiplication and gradient descent without doing so generically by using a more abstract level of description.

UC Berkeley | Software Engineer | Berkeley | Full-time, REMOTE, ONSITE, http://berkeley.edu

UC Berkeley is hiring a full-time software engineer to work on a DARPA-funded project to build next-generation methods for social science research [1]. The software engineer will join a team of scientists and engineers building Dallinger [2], an open-source platform that automates the full pipeline of crowd-sourced experimentation, from recruiting participants to managing the resultant data; think lab-on-a-chip or microfluidics, but with people. The position is a 2-year contract, renewable up to 3.5 years, and includes a comprehensive benefits package. The full job ad and instructions for applying can be found at http://bit.ly/2hCqGrj. Remote work is possible.

Our stack is Python, Flask, PostgreSQL, Redis, WebSockets, the SciPy stack, Jupyter notebooks, JavaScript, HTML, CSS, all on Heroku & AWS.

Note: if you responded to the January "Who is hiring?" post, no need to reapply — we'll be reviewing applications this week.

[1] http://www.darpa.mil/news-events/2016-03-04

[2] https://github.com/Dallinger/Dallinger

The application is back up.

There's an issue with the posting on the HR site. The university is on academic and administrative holiday until tomorrow, so this won't get fixed today, but should be back up tomorrow. Happy to answer questions here or by email in the meantime.

Looking into this now. If you email me at suchow (@berkeley.edu), I can send you an email with a working link once it's back up.

No clearance is required.

UC Berkeley | Software Engineer | Berkeley | Full-time, REMOTE, ONSITE, http://berkeley.edu

UC Berkeley is hiring a full-time software engineer to work on a DARPA-funded project to build next-generation methods for social science research [1]. The software engineer will join a team of scientists and engineers building Dallinger [2], an open-source platform that automates the full pipeline of crowd-sourced experimentation, from recruiting participants to managing the resultant data; think lab-on-a-chip or microfluidics, but with people. The position is a 2-year contract, renewable up to 3.5 years, and includes a comprehensive benefits package. The full job ad and instructions for applying can be found at http://bit.ly/2hCqGrj. Remote work is possible.

Our stack is Python, Flask, PostgreSQL, Redis, WebSockets, the SciPy stack, Jupyter notebooks, JavaScript, HTML, CSS, all on Heroku & AWS.

[1] http://www.darpa.mil/news-events/2016-03-04

[2] https://github.com/Dallinger/Dallinger

When institutions die, they take some of our memories with them. Without any cues to recall the old memory, it becomes a sort of neural dark matter that lingers within us but can never be accessed. Because people strongly tie their identities to their memories, in a way we lose a part of ourselves. Old friends and family can help here, providing throwbacks that revive memories on the brink.

Here is a well-considered riposte from Brett Glass:

https://www.quora.com/Hacker-Culture-Who-was-Brett-Glass-nam...

The premise that scientific progress has been slower in the past 50 years than in the 50 before that, and incremental in a way that it was not in the past, seems untenable to me and consistent with the authors being nostalgic for an earlier time that perhaps does not exist. I’m pretty sure someone in 1965 would be impressed to hear that (1) we can engineer viruses that allow in vivo neurons to be controlled by light; (2) we can build robots that use reinforcement learning and ML techniques to drive around a city; or (3) we discovered that humans interbred with Neanderthal, by means of sequencing 40k-year-old bones. To me these projects seem as impressive as any of the past. They all benefited from massive collaborations and would have been impossible to do alone, even by a world-class genius with a plenitude of coffee, time, and tenure.

I also object to a few more minor points. Today we have many of yesteryear’s Bell Labs and new ones spring up every year. I think we might be in the most active time period ever for independent and industrial research labs — consider for example OpenAI, the Allen Institute, RAND corporation, Charles River laboratories, Woods Hole Oceanographic Institution, Facebook, 2AI labs, Draper Laboratory, YC Research, IBM Watson Group, Communications Design Group, Xerox PARC, fivethirtyeight, New York Times R&D lab, Microsoft Research* , Google* , Bose Corporation, and SRI, not to mention the hundreds of biotech startups and tech companies that do applied reseach (*s because they are acknowledged in the article). Interestingly, I think Nokia Bell Labs might itself qualify as a modern Bell Labs — it still exists, and in 2014 a scientist was awarded the Nobel Prize for work done there in the 1980s, which is after the authors’ cutoff.

Furthermore, I’d be interested in seeing evidence that discoveries of the past can be at least in part attributed to a lack of communication between the scientist and others, and that today’s levels of communication are greater and that it has a negative impact. I have evidence that at least one world-class scientist of the past, Francis Galton, spent a ton of time communicating with others and used what he learned from that communication as a core part of his most important works (see, in particular, his autobiography, which discusses his communication in detail). Similarly, at least some classic big discoveries of the past, such as Watson & Crick’s DNA work, was possible only because there was more than one person involved — in fact, more than two.