Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!

 



Forgot your password?
typodupeerror

Comment i'm more interesed in In-memory compute (Score 1) 58

i'm currently more intereseted in nin-memory analog computing, such as with phase change memory or resistive ram memory. it gives a speedup of over 1,000x a GPGPU and consumes 1/1,000 of the power.

A big limitation of it, however, is that while it's great for implementing a trained neutal network (forward propagation), i don't see a way to get that kind of 1,000x improvement in the training / back propagation phase.

it might be possible thought. many years ago i came up with a circuit for an analog adaptive control circuit, which is kind of similiar but integrates to in memory compute but integrates the training on the same plane.

maybe people could work on circuits that combine / synthesize the two approaches; neuromorphic + in memory compute.

Comment When will Merrick Garland have his confirmation? (Score 1) 343

Can they figure when congress will schedule Merrick Garland's confirmation hearing?
It's been a few years, and the Constitution hasn't changed. It still says they have to provide advice and consent.
I checked it today, the letters didn't move around from where they were yesterday.

Comment use multiple microphones, source separation (Score 1) 144

"this new system uses a strategically placed sound meter"
They could have a more accurate and defensible fix on the vehicle doing the violation if they positions a few different microphones - maybe 4 - one at each corner -so they could use source separation algorithms, such as PCA (principal component analysist), or even more compute intensive ones like ICA, to separate the audio source, and even get a spatial fix on them, especially if they're travelling - you can add doppler effects, etc., to get a precise position and velocity of multiple sound sources.

Comment why use AI when you can solve it with algebra? (Score 1) 82

seems like a pretty simple problem with very few variables.
just build a model and fit it empirically to the data. Then run a simple optimization algorithm. If you can't solve it by taking the second derivative, just use newton's method or something.

so much less computational power. so much less computational complexity. and you get a more accurate answer.

Comment yeah, but what about math or physics lectures? (Score 1) 70

real estate appraisal or the roman empire sounds like there's not going to be a lot of complex information to retain -- just a few simple pieces of information. if it was a math lecture or a physics lecture - which requires deep attention and active, slow, careful spatial / formal operational reasoning, i'm sure the results would be very different. On that note - maybe they should compare lectures that require formal operational reasoning with lectures that only require concrete operational reasoning.

Medicine

Chinese Researchers Correct Genetic Mutation In Embryos Using Base Editing (bbc.com) 35

dryriver writes: Chinese researchers have taken tissue from a beta-thallasemia patient, created cloned embryos from that patient's cells, and used a genetic editing technique known as Base Editing to correct the gene mutation that causes beta-thallasemia. The embryos were not implanted in a womb, so no actual babies were created during the procedure. The BBC reports: "Precise 'chemical surgery' has been performed on human embryos to remove disease in a world first, Chinese researchers have told the BBC. The team at Sun Yat-sen University used a technique called base editing to correct a single error out of the three billion 'letters' of our genetic code. They altered lab-made embryos to remove the disease beta-thalassemia. The embryos were not implanted. The team says the approach may one day treat a range of inherited diseases. Base editing alters the fundamental building blocks of DNA: the four bases adenine, cytosine, guanine and thymine. Base editing works on the DNA bases themselves to convert one into another. Prof David Liu, who pioneered base editing at Harvard University, describes the approach as 'chemical surgery.' He says the technique is more efficient and has fewer unwanted side-effects than Crispr. He told the BBC: 'About two-thirds of known human genetic variants associated with disease are point mutations. So base editing has the potential to directly correct, or reproduce for research purposes, many pathogenic [mutations].'"

Comment cuts $9.2B dollars, wants credit for $.2B increase (Score 5, Insightful) 260

okay so let me get this straight, trump cuts 9.2 billion dollars from the DOE's budget, wants credit for $200M that he didn't even fund - he's just saying that of the money already allocated to grants, this much should go to STEM.

Meanwhile, when obama was president, he proposed a 4 billion dollar inceare in DOE's budget to go specifically to CS education, but that didn't pass because of republicans.

So the net score is: Obama +4 billion (blocked by republicans), Trump -9.2 billion (republicans love it).

And he wants to sell this as him supporting STEM?!?

Slashdot Top Deals

Time is nature's way of making sure that everything doesn't happen at once. Space is nature's way of making sure that everything doesn't happen to you.

Working...