Archives: Tweets
-
@moeller_ml maybe just cathartically vindictive?
@moeller_ml maybe just cathartically vindictive?
-
@psygnisfive but borrowers have “lost the keys” to…
@psygnisfive but borrowers have “lost the keys” to the rebus trick, so Aleph has no relationship to original ‘alif (ox).
-
@psygnisfive also the same rebus trick happening i…
@psygnisfive also the same rebus trick happening in early Phoenician, from which semitic, Greek, Roman alphabets derive.
-
RT @fchollet: Learn your features, learn your arch…
Learn your features, learn your architectures, learn your optimizers. Soon: write your DL papers via gradient descent.
-
This is such a concise description of PREACHER. ht…
This is such a concise description of PREACHER. https://t.co/kPWDExfeth
-
@IslamSciFi Uber will be happy to take your money…
@IslamSciFi Uber will be happy to take your money indefinitely. But many of us will not have any left to give w predatory human obsolescence
-
@scarequotes kind of a slow burn though.
@scarequotes kind of a slow burn though.
-
@scarequotes PLANETARY is still the best FF reboot…
@scarequotes PLANETARY is still the best FF reboot.
-
@shalgrim @ryageo backprop is way to learn weights…
@shalgrim @ryageo backprop is way to learn weights. architecture of backprop, hidden weights: new loci of eng effort https://t.co/pR3BbQn6Pm
-
@shalgrim @ryageo embedding is a learned feature v…
@shalgrim @ryageo embedding is a learned feature vector for each token-type minimizing some (usually NN) loss over tokens in training set.