Explainable AI: Interpreting the neuron soup of deep learning
Kate Saenko had a problem. Her AI algorithms tended to identify scientists as men and kitchen workers as women, and she didn't know why. An associate professor at Boston University's Department of Computer Science, Kate had been using deep learning to automate the captioning of images and videos. And to be true, the results were spectacular. Neural networks, the software structure that underlies deep learning, proved to be very good at generating human-like descriptions of digital imagery.
相關推薦
Explainable AI: Interpreting the neuron soup of deep learning
Kate Saenko had a problem. Her AI algorithms tended to identify scientists as men and kitchen workers as women, and she didn't know why. An associate profe
14.On the Decision Boundary of Deep Neural Networks
關於深度神經網路的決策邊界 摘要 雖然深度學習模型和技術取得了很大的經驗成功，但我們對許多方面成功來源的理解仍然非常有限。為了縮小差距，我們對訓練資料和模型進行了微弱的假設,產生深度學習架構的決策邊界。我們在理論上和經驗上證明，對於二元情形和具有常用交叉熵的多類情況，神經網路的最後權重層收斂
Journeys in big data and AI across the transport networks of London & Paris
When looking for examples of digital innovation, few of us would think of public transport. But it turns out the sector is a rich source of use cases for b
The basics of Deep Learning and Bayesian Networks in under five minutes
Still confused about deep learning, how it works, what is its shortcomings, and what is its origins? Paraphrasing Zoubin: Deep learning is neural networks
AI Is The New Face Of Systemic (And Automated) Inequality
One source of income inequality is prejudice. Unconscious (and, sadly, conscious) attitudes direct opportunities more to favored groups and steer them away
Knowledge Plus Statistics: Understanding the Emerging World of Deep Probabilistic Programming…
Knowledge Plus Statistics: Understanding the Emerging World of Deep Probabilistic Programming LanguagesThe use of statistics to overcome uncertainty is one
Theories of Deep Learning
tun topo dice stand 9.4 mar sem speed 2.0 https://stats385.github.io/readings Lecture 1 – Deep Learning Challenge. Is There Theory
李巨集毅機器學習 P13 Brief Introduction of Deep Learning 筆記
deep learning的熱度增長非常快。 下面看看deep learning的歷史。 最開始出現的是1958年的單層感知機，1969年發現單層感知機有限制，到了1980年代出現多層感知機（這和今天的深度學習已經沒有太大的區別），1986年又出現了反向傳播演算法（通常超過3
李巨集毅機器學習 P15 “Hello world” of deep learning 筆記
我們今天使用Keras來寫一個deep learning model。 tensorflow實際上是一個微分器，它的功能比較強大，但同時也不太好學。因此我們學Keras，相對容易，也有足夠的靈活性。 李教授開了一個玩笑： 下面我們來寫一個最簡單的deep learning mo
機器學習與深度學習系列連載： 第二部分 深度學習（九）Keras- “hello world” of deep learning
Keras Kearas 是深度學習小白程式碼入門的最佳工具之一。 如果想提升、練習程式碼能力，還是建議演算法徒手python實現。 複雜的深度神經網路專案還是推薦TensorFlow或者Pytorch Keras是一個高層神經網路API，Keras由純Pyt
Application of deep learning in Industrial area
Application of deep learning in Industrial area https://www.vision-systems.com/articles/print/volume-22/issue-10/departments/technology-trends/machi
Best (and Free!!) Resources to Understand Nuts and Bolts of Deep Learning
The internet is filled with tutorials to get started with Deep Learning. You can choose to get started with the superb Stanford courses CS221&nbs
A Brief Overview of Deep Learning
A Brief Overview of Deep Learning 【 注：本文是Ilya Sutskever受邀給Yisong Yue部落格寫的文章。原文在Yisong Yue部落格上：http://yyue.blogspot.com/2015/01/a-brief-o
Naftali Tishby——Information Theory of Deep Learning演講翻譯（二）
要想聽懂這一段，先準備一點基礎知識： Tishby另一個視訊，介紹的更詳細一點。 1.PAC學習：Probably Approximately Correct，PAC框架主要確定資料是否可分，確定訓練樣本個數，判斷時間空間複雜度等。 2. 假設空間：Hypoth
機器學習9：“Hello World” of deep learning
一、框架 1、TensorFlow或者theano比較靈活，可以理解成微分器，你可以用來實現Gradient Decent，但用起來實際上是有難度的； 2、keras其實是TensorFlow或theano的API介面，可以幫助你快速做一個模型，現在TensorFlow預設使用Keras介面
Essentials of Deep Learning: Visualizing Convolutional Neural Networks in Python
Introduction One of the most debated topics in deep learning is how to interpret and understand a trained model – particularly in the con
Bridging the Deployment Gap for Deep Learning (Part 2)
From Exploration to Production — Bridging the Deployment Gap for Deep Learning (Part 2)This is the second part of a series of two blogposts on deep learnin
How to rapidly test dozens of deep learning models in Python
Although k-fold cross validation is a great way of assessing a model’s performance, it’s computationally expensive to obtain these results. We can simply s
Competitive Advantages of Deep Learning for Your Business
What do you think of when you hear about AI? Do you picture your favourite sci-fi movie or a book that you read when you were younger? In that favourite bo
Plant-Disease-Recognition App Presented on the Final Day of Yantra Learning 2017/18
As part of Yantra Learning 2017/18, Nepal’s first machine learning competition, Team RARS from IOE, Pulchowk campus presented their final demo on the final