How long does it take to get access to AutoML?
Using Google Cloud AutoML to Classify Poisonous Australian Spiders
31–40 of 42 posts
Re: Using Google Cloud AutoML to Classify Poisonous Australian Spiders
#32I like the content of this blog post, but the use of stock photos was really off-putting and I had to almost force myself to keep on reading.
Re: Using Google Cloud AutoML to Classify Poisonous Australian Spiders
#33Something like this could actually save lives. Take a photo of what bit you, then the app could provide an answer (with a confidence level) of "do I need to go to the hospital" and provide a few example apps for various matches with level of danger. Thanks for sharing!
I don't think it is actually likely to save lives, because spiders very rarely kill. As far as I can tell, there's been one death from a spider bite in Australia in 40 years: https://www.telegraph.co.uk/news/2016/04/12/young-man-dies-a... .
You'd be crazy not to get a second opinion at a hospital though right? "The app said I didn't need to see a doctor" sounds like dangerous territory.
Not saying a doctor is perfect, though I find they're often better than they get credit for.
Re: Using Google Cloud AutoML to Classify Poisonous Australian Spiders
#34Earlier quoted context omitted.
There's no such thing as auto-labeling. Data Turks is manual labeling. There is active learning[1] and related algorithms where you trace the boundary of your classifier and pass examples along that boundary to be manually labeled (as they are the ones the classifier is most unsure about). But there is nothing "auto" about this - it's just being smart about where to deploy the manual labor. [1] https://en.wikipedia.o…
Lets say we want to create a labeled data for text summarization for medium articles. Could the highlighted part be used as summary, its not auto labeled per se, but can be a proxy and passed to labelers to verify/edit.
Sure. There are lots of useful proxies for labeled data.
It's worth noting that highlighted sections in Medium articles probably aren't great summaries (they are more a representation of important points - which is a useful thing to predict as well).
For example, many summarizer systems are trained on the single-line summaries given in news media systems. There have been attempts to use Tweets as summaries for linked articles too.
Re: Using Google Cloud AutoML to Classify Poisonous Australian Spiders
#35Earlier quoted context omitted.
Generally, a redback is fairly unpleasant, but doesn't require a trip to the hospital (for a healthy adult), so its unlikely to be fatal in most cases anyway (thankfully!). Funnel-webs are terrifying.
A redback bite generally requires medical attention; though not triple-0 level urgency, it shouldn't be a wait-and-see-how-you-feel treatment.
Unless you're a kid, and then even the antidote is terrifying.
Re: Using Google Cloud AutoML to Classify Poisonous Australian Spiders
#36I'm surprised someone is yet to point out the error of using the word poisonous instead of venomous. I suppose it's not important to the article.
Although, if you do decide to nibble one you may find it is indeed disagreeable.
Re: Using Google Cloud AutoML to Classify Poisonous Australian Spiders
#37Is it possible to export the trained model from AutoML for offline use?
They haven't come to a decision as to whether that will be offered AFAIK. I really hope they decide to allow downloading the model.
Re: Using Google Cloud AutoML to Classify Poisonous Australian Spiders
#38This is huge, and it's only an alpha. I begun reading about AutoML/Neural Architecture searches around ~year ago and something I've been thinking about is: Why doesn't this just move the optimization problem? Aren't you now just optimizing your DeepRL network rather than the network you're trying to optimize?
The idea of AutoML (in this case[1]) is to improve the NN architecture for a given type of problem. In "normal" machine learning this is basically hyperparmater optimization for a given dataset (eg, the depth of a random forest, XGB parameters, the best random seed/jk ) In this case is tests different combinations of operators on a known dataset to see what performs the best. So it is optimizing the prediction networ…
Jeff Dean talks about AutoML using RL and in the paper "
Neural Architecture Search with Reinforcement Learning" it also talks about this.
Also it seems different from more traditional hyperparameter optimization because it makes novel cells. So the structure of the network isn't limited to our existing library of layers/cells.
https://arxiv.org/abs/1611.01578 https://youtu.be/HcStlHGpjN8?t=2073
Re: Using Google Cloud AutoML to Classify Poisonous Australian Spiders
#39Earlier quoted context omitted.
The idea of AutoML (in this case[1]) is to improve the NN architecture for a given type of problem. In "normal" machine learning this is basically hyperparmater optimization for a given dataset (eg, the depth of a random forest, XGB parameters, the best random seed/jk ) In this case is tests different combinations of operators on a known dataset to see what performs the best. So it is optimizing the prediction networ…
Jeff Dean talks about AutoML using RL and in the paper " Neural Architecture Search with Reinforcement Learning" it also talks about this. Also it seems different from more traditional hyperparameter optimization because it makes novel cells. So the structure of the network isn't limited to our existing library of layers/cells. https://arxiv.org/abs/1611.01578 https://youtu.be/HcStlHGpjN8?t=2073
"Novel Cells" are combinations of existing operators.
It's entirely true that these are combinations that humans haven't (and probably wouldn't) come up with.
I don't want to underplay this. "It's similar to hyperparameter search" makes it sound like it isn't interesting or novel, which is untrue. I completely believe it is a revolutionary way to build software (so much so that I quit my job, raised funding and are working on a similar space of problems).
But it isn't doing something like inventing a new math operations similar to the other operators which humans put together to form cells/layers. It is rearranging and choosing those operators in new ways.
Re: Using Google Cloud AutoML to Classify Poisonous Australian Spiders
#40Earlier quoted context omitted.
Jeff Dean talks about AutoML using RL and in the paper " Neural Architecture Search with Reinforcement Learning" it also talks about this. Also it seems different from more traditional hyperparameter optimization because it makes novel cells. So the structure of the network isn't limited to our existing library of layers/cells. https://arxiv.org/abs/1611.01578 https://youtu.be/HcStlHGpjN8?t=2073
"Novel Cells" are combinations of existing operators. It's entirely true that these are combinations that humans haven't (and probably wouldn't) come up with. I don't want to underplay this. "It's similar to hyperparameter search" makes it sound like it isn't interesting or novel, which is untrue. I completely believe it is a revolutionary way to build software (so much so that I quit my job, raised funding and are w…
Okay I see what you're saying and I completely agree.