Stealing Machine Learning Models via Prediction APIs · HackerPapers discussions

Stealing Machine Learning Models via Prediction APIs — HN submission

Sun, 20 Jan 2019 08:41:54 GMT

Stealing Machine Learning Models via Prediction APIs (2016) [pdf]

3 points; 0 discussion replies on HN. Paper and source history. A submission is not an endorsement.

Stealing Machine Learning Models via Prediction APIs — mentioned by taurine

Wed, 13 Jun 2018 16:13:47 GMT

It is referring to https://arxiv.org/abs/1609.02943 What you are referring to is possible, but is not "copying" per se, just trying to infer what the system is doing (inverse RL), and then exploit that/make it do mistakes. If you are not HFT it is very difficult to distinguish bots from humans, so you'd have a hard time even finding a target.

Paper and discussion history. A mention is not an endorsement.

Stealing Machine Learning Models via Prediction APIs — HN submission

Thu, 19 Apr 2018 21:25:34 GMT

Stealing Machine Learning Models via Prediction APIs (2016)

2 points; 0 discussion replies on HN. Paper and source history. A submission is not an endorsement.

Stealing Machine Learning Models via Prediction APIs — mentioned by red_dinosaur

Tue, 28 Mar 2017 01:49:54 GMT

The author of this story clearly does not know what "reverse engineer" means. This is reverse engineering AI companies: https://arxiv.org/abs/1609.02943

Paper and discussion history. A mention is not an endorsement.

Stealing Machine Learning Models via Prediction APIs — mentioned by zek

Wed, 08 Mar 2017 22:06:25 GMT

Yes! there is an interesting paper on this subject: https://arxiv.org/abs/1609.02943

Paper and discussion history. A mention is not an endorsement.

Stealing Machine Learning Models via Prediction APIs — mentioned by Jach

Wed, 08 Mar 2017 21:15:30 GMT

How long until someone runs an attack like https://arxiv.org/abs/1609.02943 and provides the model for free?

Paper and discussion history. A mention is not an endorsement.

Stealing Machine Learning Models via Prediction APIs — mentioned by __coaxialcabal

Wed, 18 Jan 2017 13:39:14 GMT

I'm not sure if open sourcing the data is absolutely necessary to reproduce the results given access to the prediction API. "Stealing Machine Learning Models via Prediction APIs" (https://arxiv.org/abs/1609.02943) examines techniques in this space. The API might also be used in adversarial training. The underlying data would be useful for advancing research, but not necessarily a prerequisite for reproducing the API. AWS wins not because they have created an impenetrable black box, but because the economics of reverse engineering (and maintaining) are not cost effective for the majority of…

Paper and discussion history. A mention is not an endorsement.

Stealing Machine Learning Models via Prediction APIs — mentioned by grzm

Fri, 30 Dec 2016 09:57:12 GMT

Odd, isn't it? Maybe a copy/paste error? There's a paper from Oct 2016 on arXiv with this title: https://arxiv.org/abs/1609.02943

Paper and discussion history. A mention is not an endorsement.

Stealing Machine Learning Models via Prediction APIs — mentioned by seycombi

Sat, 24 Dec 2016 18:45:08 GMT

Mentioned paper "Stealing Machine Learning Models via Prediction APIs" https://arxiv.org/abs/1609.02943

Paper and discussion history. A mention is not an endorsement.

Stealing Machine Learning Models via Prediction APIs — HN submission

Thu, 22 Sep 2016 16:00:28 GMT

Stealing Machine Learning Models via Prediction APIs

206 points; 37 discussion replies on HN. Paper and source history. A submission is not an endorsement.