2020-08-29 16:58:38 In our previous collaboration's I've benefited a lot from using David's machine learning frameworks. If you're a researcher or student looking for both simplicity and customizability, definitely check out Objax. https://t.co/si8glTP5CQ
2020-08-28 01:53:41 RT @ShreyaR: Our AI/ML residency program is here!! I'll be a resident host on @goodfellow_ian's team at SPG. Please apply, or spread the…
2020-08-27 22:38:57 @nuttyrahul Yes, do apply.
2020-08-27 22:38:03 @chriscanal4 Yes, they'll get to work with me personally. Remote work is possible in the short term but the expectation is that when COVID conditions are better this will be a Santa Clara position.
2020-08-27 22:35:02 @vandersonmr_ My group isn't taking interns right now but there are a lot of other great ML internship opportunities around Apple and I'm generally available to chat with interns from other teams
2020-08-27 22:31:40 Apple now has an AI/ML residency program! I'm looking forward to working with our first class of residents. https://t.co/KcQnPbpt2f
2020-08-25 00:07:50 @jaschasd @pfau @negative_result @duck @sschoenholz @ethansdyer @dwf If you train an fc layer with conv constraints, then continue training with the constraints removed, does sgd actually make the model get worse? Or just can’t find the good solution without constraints but preserves it once found?
2020-08-25 00:06:03 @jaschasd @pfau @negative_result @duck @sschoenholz @ethansdyer @dwf That’s very interesting. You can think of a conv layer as just fc with linear constraints during optimization. Weird that some information is accessible to constrained optimization but not to unconstrained.
2020-08-24 01:42:51 @jaschasd @pfau @negative_result @duck @sschoenholz @ethansdyer For the maxout paper, @dwf got a SOTA on CIFAR10, using ZCA whitening. But the ZCA was regularized by adding a constant to the diagonal of the covariance matrix. Is that enough to be out of the scope of your claim?
2020-08-29 16:58:38 In our previous collaboration's I've benefited a lot from using David's machine learning frameworks. If you're a researcher or student looking for both simplicity and customizability, definitely check out Objax. https://t.co/si8glTP5CQ
2020-08-28 01:53:41 RT @ShreyaR: Our AI/ML residency program is here!! I'll be a resident host on @goodfellow_ian's team at SPG. Please apply, or spread the…
2020-08-27 22:38:57 @nuttyrahul Yes, do apply.
2020-08-27 22:38:03 @chriscanal4 Yes, they'll get to work with me personally. Remote work is possible in the short term but the expectation is that when COVID conditions are better this will be a Santa Clara position.
2020-08-27 22:35:02 @vandersonmr_ My group isn't taking interns right now but there are a lot of other great ML internship opportunities around Apple and I'm generally available to chat with interns from other teams
2020-08-27 22:31:40 Apple now has an AI/ML residency program! I'm looking forward to working with our first class of residents. https://t.co/KcQnPbpt2f
2020-08-25 21:56:37 My team is hiring ML/RL infra engineers: https://t.co/JqxwOrJziC
2020-08-25 00:07:50 @jaschasd @pfau @negative_result @duck @sschoenholz @ethansdyer @dwf If you train an fc layer with conv constraints, then continue training with the constraints removed, does sgd actually make the model get worse? Or just can’t find the good solution without constraints but preserves it once found?
2020-08-25 00:06:03 @jaschasd @pfau @negative_result @duck @sschoenholz @ethansdyer @dwf That’s very interesting. You can think of a conv layer as just fc with linear constraints during optimization. Weird that some information is accessible to constrained optimization but not to unconstrained.
2020-08-24 01:42:51 @jaschasd @pfau @negative_result @duck @sschoenholz @ethansdyer For the maxout paper, @dwf got a SOTA on CIFAR10, using ZCA whitening. But the ZCA was regularized by adding a constant to the diagonal of the covariance matrix. Is that enough to be out of the scope of your claim?