登录查看更多内容

Attacking machine learning with adversarial examples

Diego Marinho de Oliveira

Gen-AI Search, RecSys | ex-SEEK, AI Lead, Data Scientist Manager and ML Engineer Specialist

发布日期: 2017年2月21日

Abstract Adversarial examples are inputs to machine learning models that an attacker has intentionally designed to cause the model to make a mistake; they're like optical illusions for machines. In this post we'll show how adversarial examples work across different mediums, and will discuss why securing systems against them can be difficult.

Authors: IAN GOODFELLOW, NICOLAS PAPERNOT, SANDY HUANG, YAN DUAN, PIETER ABBEEL, JACK CLARK

Read full article at https://bit.ly/2lTj1nO

Monika Kovacs

///

8 年

Very interesting, thank you. Gradient masking looks a little scary.

1 次回应

要查看或添加评论，请登录

Diego Marinho de Oliveira的更多文章

Deep Learning for Personalized Search and Recommender Systems

2017年10月5日

Deep Learning for Personalized Search and Recommender Systems

Nice review about Deep Learning for Search + Recommender Systems: "Abstract Deep learning has been widely successful in…

2 条评论
Facets: An Open Source Visualization Tool for Machine Learning Training Data

2017年7月18日

Facets: An Open Source Visualization Tool for Machine Learning Training Data

"Abstract Getting the best results out of a machine learning (ML) model requires that you truly understand your data…

1 条评论
Spark: The Definitive Guide

2017年7月6日

Spark: The Definitive Guide

Databricks published some free chapters today about Spark. "Apache Spark has seen immense growth over the past several…

5 条评论
One Model To Learn Them All

2017年6月27日

One Model To Learn Them All

"Abstract Deep learning yields great results across many fields, from speech recognition, image classification, to…

3 条评论
Do Balancing Classes Improve Classifier Performance?

2017年5月25日

Do Balancing Classes Improve Classifier Performance?

Nice post by Nina Zumel "It’s a folk theorem I sometimes hear from colleagues and clients: that you must balance the…

4 条评论
Neural Audio Synthesis of Musical Notes with WaveNet Autoencoders

2017年5月21日

Neural Audio Synthesis of Musical Notes with WaveNet Autoencoders

"Abstract Generative models in vision have seen rapid progress due to algorithmic improvements and the availability of…
Neural Ranking Models with Weak Supervision

2017年5月5日

Neural Ranking Models with Weak Supervision

Abstract Despite the impressive improvements achieved by unsupervised deep neural networks in computer vision and NLP…
Simple/Incomplete Benchmark of Machine Learning Libraries for Classification

2017年4月19日

Simple/Incomplete Benchmark of Machine Learning Libraries for Classification

The sharing for today :) All benchmarks are wrong, but some are useful "This project aims at a minimal benchmark for…

1 条评论
Introducing tf-seq2seq: An Open Source Sequence-to-Sequence Framework in TensorFlow

2017年4月12日

Introducing tf-seq2seq: An Open Source Sequence-to-Sequence Framework in TensorFlow

Summary: "In addition to machine translation, tf-seq2seq can also be applied to any other sequence-to-sequence task…

1 条评论
Mask R-CNN

2017年3月22日

Mask R-CNN

"Abstract We present a conceptually simple, flexible, and general framework for object instance segmentation. Our…

3 条评论

See all articles

Attacking machine learning with adversarial examples

Diego Marinho de Oliveira

Gen-AI Search, RecSys | ex-SEEK, AI Lead, Data Scientist Manager and ML Engineer Specialist

Diego Marinho de Oliveira的更多文章

社区洞察

其他会员也浏览了

An activation function helps machines think beyond straight lines to understand complex patterns.

k-means & k-nearest Neighbours in ml

“Formal Verification Made Simple” Series | Article-3 | Simplifying Model Checking

Forest Fire Detection using CNN - Part 2

GPTQ Quantization of Poro-34B LoRA fine-tuned LLM with S Group data

ENSEMBLE MODELS #aiNuggets #MLBytes

A multi-layered support vectors machine

Porting machine learning models to mobile for Edge inference

Learning from Little: Comparison of Classifiers Given Little Training

TuckER: Tensor Factorization for Knowledge Graph Completion

Diego Marinho de Oliveira的更多文章

Deep Learning for Personalized Search and Recommender Systems

Facets: An Open Source Visualization Tool for Machine Learning Training Data

Spark: The Definitive Guide

One Model To Learn Them All

Do Balancing Classes Improve Classifier Performance?

Neural Audio Synthesis of Musical Notes with WaveNet Autoencoders

Neural Ranking Models with Weak Supervision

Simple/Incomplete Benchmark of Machine Learning Libraries for Classification

Introducing tf-seq2seq: An Open Source Sequence-to-Sequence Framework in TensorFlow

Mask R-CNN

社区洞察

其他会员也浏览了

An activation function helps machines think beyond straight lines to understand complex patterns.

k-means & k-nearest Neighbours in ml

“Formal Verification Made Simple” Series | Article-3 | Simplifying Model Checking

Forest Fire Detection using CNN - Part 2

GPTQ Quantization of Poro-34B LoRA fine-tuned LLM with S Group data

ENSEMBLE MODELS #aiNuggets #MLBytes

A multi-layered support vectors machine

Porting machine learning models to mobile for Edge inference

Learning from Little: Comparison of Classifiers Given Little Training

TuckER: Tensor Factorization for Knowledge Graph Completion