logane
Joined 11 years ago
Joined 11 years ago
In software we have a nice compilation process that transforms code into machine code. However, to "compile" an ASIC you go from a hardware description language like verilog (which basically describes the design) and pass it through a complicated pipeline that composes a number of different tools. Right now, engineers use ad-hoc flows that use bash/TCL to glue together all the parts --- the project posted above is an attempt to cleanly specify + control the "compilation" process.
For (a), maybe surprisingly the answer is mostly yes! Better ImageNet accuracy generally corresponds to better out of distribution accuracy. For (b), it turns out that the ImageNet dataset is full of contradictions---many images have multiple ImageNet-relevant objects, and often are ambiguously or mis-labeled, etc---so it's hard to disentangle progress in identifying objects vs. models overfitting to the quirks of the benchmark.
[1] ObjectNet: https://objectnet.dev / associated paper
[2] ImageNet-v2: https://arxiv.org/abs/1902.10811
[3] An Unbiased Lookat Dataset Bias: https://people.csail.mit.edu/torralba/publications/datasets_... (pre-AlexNet!)
[4] From ImageNet to Image Classification: https://arxiv.org/abs/2005.11295
[5] Are we done with ImageNet? https://arxiv.org/abs/2006.07159
[6] Evaluating Machine Accuracy on ImageNet: http://proceedings.mlr.press/v119/shankar20c.html
The mirror system isn't quite as macgyvered though :)
Do you have any evidence of this happening? I've never heard of pro-US censorship before on any of these platforms (or any social media platform..)
http://loganengstrom.com/