AI News, What is a hacker's approach to learning machine learning and computer vision fast without much math?

What is a hacker's approach to learning machine learning and computer vision fast without much math?

Broadly there are three possible categories: (1) Straightforward and a potentially limited system : If you just wish to implement a very straightforward system (train a standard CNN on a given dataset, do SVM classification, do basic segmentation, face detection, etc.) on a platform that provides enough resources for the algorithms to run, and no one is going to ever ask you to improve the performance, you do not need to learn any math, and thus no background behind ML and Vision Algorithms.

For instance, if one asks you that he needs you to improve the performance on a given dataset, you might then need to use bias-variance trade-off concepts, concepts of empirical risk minimization and true error (under an expected sense), understanding why an algorithm works, and inferring from the mathematics of the algorithm what is missing inside.

Also one should try to always analyse ML and vision algorithms from the following perspectives : (a) where the algorithms can be robust and why (b) under what conditions will the algorithms fail and why (c) what is the computational cost of running the algorithm and is it worth the trade off between computational efficiency and slightly improved (not perceivable enough in an application) performance.

CppCon 2014: Chandler Carruth "Efficiency with Algorithms, Performance with Data Structures"

— Presentation Slides, PDFs, Source Code and other presenter materials are available at: .

NIPS 2011 Big Learning - Algorithms, Systems, & Tools Workshop: High-Performance Computing...

Big Learning Workshop: Algorithms, Systems, and Tools for Learning at Scale at NIPS 2011 Invited Talk: High-Performance Computing Needs Machine ...

NIPS 2011 Big Learning - Algorithms, Systems, & Tools Workshop: Hazy - Making Data-driven...

Big Learning Workshop: Algorithms, Systems, and Tools for Learning at Scale at NIPS 2011 Invited Talk: Hazy: Making Data-driven Statistical Applications ...

CppCon 2016: Chandler Carruth “High Performance Code 201: Hybrid Data Structures"

— Presentation Slides, PDFs, Source Code and other presenter materials are available at: — Modern .

Designing for Understandability: The Raft Consensus Algorithm

This talk was presented by Professor John Ousterhout on August 29, 2016 as part of the CS @ Illinois Distinguished Lecture Series. It is shared here with his ...

Lecture 14 | Deep Reinforcement Learning

In Lecture 14 we move from supervised learning to reinforcement learning (RL), in which an agent must learn to interact with an environment in order to ...

Algorithms (Time Series Segmentation) | Medical Data Mining L01T05 | Introduction & Scientific Know.

The Online Certificate Program in Genomics and Biomedical Informatics Bar-Ilan University & Sheba Medical Center Course 803.80-675 - Medical Data Mining ...

Stop assuming data, algorithms and AI are objective | Mata Haggis-Burridge | TEDxDelft

Mata Haggis-Burridge works in the gaming industry, where he works with Artificial Intelligince systems every day. Too often, he sees that people assume AI is ...

Context Matters: Distributed Graph Algorithms and Runtime Systems

In this video from the PASC16 conference, Andrew Lumsdaine from Indiana University presents: Context Matters: Distributed Graph Algorithms and Runtime ...

CppCon 2017: Carl Cook “When a Microsecond Is an Eternity: High Performance Trading Systems in C++”

— Presentation Slides, PDFs, Source Code and other presenter materials are available at: .