Industry
End-to-End Weak Supervision Carnegie Mellon University 2
Aggregating multiple sources of weak supervision (WS) can ease the data-labeling bottleneck prevalent in many machine learning applications, by replacing the tedious manual collection of ground truth labels. Current state of the art approaches that do not use any labeled training data, however, require two separate modeling steps: Learning a probabilistic latent variable model based on the WS sources - making assumptions that rarely hold in practice - followed by downstream model training. Importantly, the first step of modeling does not consider the performance of the downstream model. To address these caveats we propose an end-to-end approach for directly learning the downstream model by maximizing its agreement with probabilistic labels generated by reparameterizing prior probabilistic posteriors with a neural network. Our results show improved performance over prior work in terms of end model performance on downstream test sets, as well as in terms of improved robustness to dependencies among weak supervision sources.
Hotline Miami meets football, the power of video editing and other new indie games worth checking out
Valve's Steam Machine: Everything we know Welcome to our latest roundup of what's going on in the indie game space. As always, we've got a bunch of neat games to tell you about. Perhaps I'll tear myself away from playing as Chappell Roan in or Jetpack Cat in long enough to check more of them out. Thanks to the folks at, I learned about a short, text-based game from Woe Industries from a while back called . Surprisingly enough, you take on the role of a venture capitalist who has plowed gobs of money into genAI technology and might be starting to have doubts about that investment.