praccu

I built my own motion activated stairwell lights using an ESP32 and Home Assistant. Works pretty well.

So then I decided to set up some LED strip lights for my office with really nice wireless buttons from IKEA.

I confirmed that the buttons and the lights both were zigbee compatible, but I missed that one supported only Thread and the other only Matter. So now I have two identical antennas, each handling a different protocol.

Cool stuff. I first worked on ML for exploratory synthesis in 2012, and am still in related areas.

Once you have the experimental loop running, I suspect it will be quite difficult to hill climb on this task.

There will be some improvements you can make to the harness, but I suspect you'll be doing a lot of human in the loop review and providing feedback that goes back into the harness instructions.

I know it's fashionable to imagine automating the whole process, but everything I've seen is that the only systems that succeed are the ones that are augmenting an expert.

This idea seems to be coming up in multiple places.

Here's one from Deepmind:

https://arxiv.org/abs/2509.10147

Many years ago (13?), I was around when Amazon moved SABLE from RAM to SSDs. A whole rack came from a single batch, and something like 128 disks went out at once.

I was an intern but everyone seemed very stressed.

https://www.nature.com/articles/s41586-024-07744-y https://research.google/blog/fast-accurate-climate-modeling-...

You're very confused.

https://www.census.gov/newsroom/stories/poverty-awareness-mo...

I should update the post with details.

I ended up using DFRobot firebeetle because it respects sleep mode. I'm at 15 months on one charge with a 10Ah battery.

E Ink Todo list on my wall that pulls from Todoist, which I can update from my watch.

https://blog.praccu.com/

Massachusetts?

Not for sale, I'm hijacking your thread, but I wear the heck out of my Google+ slippers.

Shameless self promotion: I wrote one of the more cited papers in the field [0], back in 2016.

A key challenge: very few labs have enough data.

Something I view as a key insight: a lot of labs are doing absurdly labor intensive exploratory synthesis without clear hypotheses guiding their work. One of our more useful tasks turned out to be interactively helping scientists refine their experiments before running them.

Another was helping scientists develop hypotheses for _why_ reactions were occuring, because they hadn't been able to build principled models that predicted which properties were predictive of reaction formation.

Going all the way to synthesis is nice, but there's a lot of lower hanging fruit involved in making scientists more effective.

[0] https://www.nature.com/articles/nature17439

At Amazon I set up an evaluation approach based on whether the system completed the desired task (in that context it was "did the search result using the speech recognition return the same set of items to buy as the transcript.)

https://scholar.google.com/citations?view_op=view_citation&h...

Shameless plug: I published almost the same thing in a very nearby field 2 years ago:

https://www.nature.com/articles/nature17439

TLDR: DWave doesn't provide a speed-up over classical computers.

http://www.scottaaronson.com/blog/?p=3192

I'm not with the lab any more, but last I heard they were soliciting data from nearby but related subfields (our lab was doing mostly vanadium templated tellurides, but there are a lot of other things to explore in the space of inorganic–organic hybrids, and there are labs working on them). The problem really is data sharing; the most active collaborators were willing to give us lab note-books as long as we paid to digitize them; it was very hard to convince other labs to digitize their own notebooks. TBH I think that's the hard part of a lot of this (collecting the data).

We did something pretty similar in a nearby field of materials science, using support vector machines rather than linear regression and focusing on a less specific target:

http://www.nature.com/nature/journal/v533/n7601/abs/nature17...

LSTMs are very different. Think of it this way. LSTMs store information about the current problem you're solving. DNCs store information about the world.

LSTMs are designed to capture long range dependencies, e.g., "this word at the start of the sentence interacts with this word at the end of the sentence."

DNCs are designed to incorporate outside information, e.g., "i happen to know (from background knowledge) that these two people in this sentence are married"

As an alternative perspective: I worked in a lab that was relying heavily on intuition and heuristics for exploratory synthesis. By digitizing all of the lab's data and doing some straightforward machine learning, we were able to sufficiently improve the efficiency of the lab, compared to using intuition, and help the chemists develop a deeper understanding of the system they were studying:

http://www.nature.com/news/computer-gleans-chemical-insight-...

So, I guess my question is: to what degree is this reliance on intuition really just a lack of good data driven methods?

It's really true, though. It's brutal, ugly work with no hope of an end.

Edit: Favorite paper on the topic: http://research.google.com/pubs/pub43146.html