I have a major love/hate affair with google. This MindBlog uses google's blogger service, and so utterly depends on it, all my email addresses forward to my gmail account, I use it to synchronize all my calendar, documents, spreadsheets, and contacts, across multiple devices. I use google voice for phoning, google+ hangouts for video chats, etc. etc. Google's services have become such a prosthesis for me that I am quite helpless away from its Cloud. At the same time, I resist as many of the 'connectivity' efforts as much as I can. I emphatically do not want to know whether a friend is nearby, and don't want people following me. I think we are constantly flirting with the 'uncanny valley' effect, where what might be useful suddenly becomes very spooky.
Here is the abstract from Le et al.(PDF here):
We consider the problem of building high-level, class-specific feature detectors from only unlabeled data. For example, s it possible to learn a face detector using only unlabeled images? To answer this, we train a 9- layered locally connected sparse autoencoder with pooling and local contrast normalization on a large dataset of images (the model has 1 billion connections, the dataset has 10 million 200x200 pixel images downloaded from the Internet). We train this network using model parallelism and asynchronous SGD on a cluster with 1,000 machines (16,000 cores) for three days. Contrary to what appears to be a widely-held intuition, our experimental results reveal that it is possible to train a face detector without having to label images as containing a face or not. Control experiments show that this feature detector is robust not only to translation but also to scaling and out-of-plane rotation. We also found that the same network is sensitive to other high-level concepts such as cat faces and human bodies. Starting with these learned features, we trained our network to obtain 15.8% accuracy in recognizing 20,000 object categories from ImageNet, a leap of 70% relative improvement over the previous state-of-the-art.