The Next Generation of Neural Networks

Google Tech Talks November, 29 2007 In the 1980’s, new learning algorithms for neural networks promised to solve difficult classification tasks, like speech or object recognition, by learning many layers of non-linear features. The results were disappointing for two reasons: There was never enough labeled data to learn millions of complicated features and the learning was much too slow in deep neural networks with many layers of features. These problems can now be overcome by learning one layer of features at a time and by changing the goal of learning. Instead of trying to predict the labels, the learning algorithm tries to create a generative model that produces data which looks just like the unlabeled training data. These new neural networks outperform other machine learning methods when labeled data is scarce but unlabeled data is plentiful. An application to very fast document retrieval will be described. Speaker: Geoffrey Hinton Geoffrey Hinton received his BA in experimental psychology from Cambridge in 1970 and his PhD in Artificial Intelligence from Edinburgh in 1978. He did postdoctoral work at Sussex University and the University of California San Diego and spent five years as a faculty member in the Computer Science department at Carnegie-Mellon University. He then became a fellow of the Canadian Institute for Advanced Research and moved to the Department of Computer Science at the University of Toronto. He spent three years from 1998 until 2001 setting up the

15 thoughts on “The Next Generation of Neural Networks

  1. why should I kill myself wtf? did you misunderstand me at some point or what? I have full head of THICK hair and a penis obviously you are dreaming about. if anybody i want baldys to die out by not having children. now shut up or i will slap your brains out.

  2. because i have a mental image of you frantically staring at an ICD Code then back at your barely passable weenis, and a gigantic sigh of relief when you realise you don’t need to hang yourself due to your backwards attitude to inferior genes… yet

  3. i just think the irony is too great being a 30-year old man with nothing else better to do than go on computer science videos he is seemingly totally disinterested in the people who are likely far more successful and respected than him, and try to abuse them on standards only he thinks is important. i mean the whole thing smacks of tiny cock syndrome and it is just too dumb not to be a joke

  4. Some of them know, of course. And they also hate baldys. It’s disturbing – the sight of this disease and the knowledge that probably those people are spreading this disease to the children of the next generation. And noteworthy I have no BALD acquaintances. I avoid them at all costs, I can’t picture myself hanging out with a bald faggot. I would probably SLAP HIS BRAINS OUT at some point lol.

  5. do people in your family or social circles know you dislike bald people this much? because honestly, nobody else really cares and just think you’re a bit weird and pointless as a person.

  6. this is a really nice, gentle introduction to neural networks. there’s a good reason he’s one of the most popular machine learning intellectuals out there

  7. there are strict criteria for that. if it’s really really small then yes, it’s called hypoplasia of penis – Q55.62 in ICD-10. criteria for diseases are strictly defined and something that is out of normal range (like ZERO hair on top of your head, or 4 cm penis) will always be a disease, genetic or otherwise.

Leave a Reply