Demystifying Machine Learning
2 likes279 views
This presentation was covered as part of Divum's New Product Developers Meet held on Apr-8th, 2017. Idea of this presentation is to gently introduce machine learning thinking and expose the tools & options available to get started. This also introduces the Google TensorFlow, Amazon ML & other ML APIs.
Convert to study materialsBETA
Transform any presentation into ready-made study materialselect from outputs like summaries, definitions, and practice questions.
![Programmatic Example - TensorFlow
import numpy as np
import tensorflow as tf
# Model parameters
W = tf.Variable([.3], tf.float32)
b = tf.Variable([-.3], tf.float32)
# Model input and output
x = tf.placeholder(tf.float32)
linear_model = W * x + b
Data Variables
& Equations](https://image.slidesharecdn.com/divumnpdmachinelearning-170411165357/85/Demystifying-Machine-Learning-14-320.jpg)
![Programmatic Example - TensorFlow
print(sess.run(linear_model, {x:[1,2,3,4]}))
y = tf.placeholder(tf.float32)
squared_deltas = tf.square(linear_model - y)
loss = tf.reduce_sum(squared_deltas)
print(sess.run(loss, {x:[1,2,3,4], y:[0,-1,-2,-3]}))
fixW = tf.assign(W, [-1.])
fixb = tf.assign(b, [1.])
sess.run([fixW, fixb])
print(sess.run(loss, {x:[1,2,3,4], y:[0,-1,-2,-3]}))
Data
& Run](https://image.slidesharecdn.com/divumnpdmachinelearning-170411165357/85/Demystifying-Machine-Learning-15-320.jpg)
Demystifying Machine Learning
- 2. Data Models Training Predictions Numbers Equations Solution Builder Apply Solutions ML Thinking
- 3. Data Numbers ML Thinking - Data Real Estate Price Estimator #Area #Sq.Ft Good enough? #Bedrooms #Years built
- 4. Models Equations ML Thinking - Models ? def estimate_house_sales_price(num_of_bedrooms, sqft, neighborhood): price = 0 # In my area, the average house costs $200 per sqft price_per_sqft = 2000 if neighborhood == "domlur": # but some areas cost a bit more price_per_sqft = 4000 elif neighborhood == "jayanagar": # and some areas cost less price_per_sqft = 10000 # start with a base price estimate based on how big the place is price = price_per_sqft * sqft # now adjust our estimate based on the number of bedrooms if num_of_bedrooms == 0: # Studio apartments are cheap price = price200000 else: # places with more bedrooms are usually # more valuable price = price + (num_of_bedrooms * 10000) return price
- 5. Models Equations ML Thinking - Models
- 6. ML Thinking - Training Training Solution Builder Supervised Training
- 7. ML Thinking - Training Training Solution Builder Unsupervised Training Bedrooms Sq.feet Area 2 2000 Domlur 3 3000 Jayanagar 1 1800 NT 4 3000 JP Nagar Price ?
- 8. ML Thinking - Training Training Solution Builder Bedrooms Sq.feet Area Sale Price My Guess 2 2000 Domlur 40L 38L 3 3000 Jayanagar 80L 90L 1 1800 NT 40L 30L 4 3000 JP Nagar 65L 86L
- 9. ML Thinking - Training Training Solution Builder
- 10. ML Thinking - Training Training Solution Builder
- 11. ML Thinking - Training Predictions Apply Solutions Bedrooms Sq.feet Area Sale Price Accuracy 2 2000 Domlur 39.9L 99%
- 12. A visual introduction to machine learning http://www.r2d3.us/visual-intro-to- machine-learning-part-1/ Visual Example from bigml.com
- 13. Programmatic Example - TensorFlow The central unit of data in TensorFlow is the tensor A tensor consists of a set of primitive values shaped into an array of any number of dimensions. TensorFlow provides multiple APIs (in Python)
- 14. Programmatic Example - TensorFlow import numpy as np import tensorflow as tf # Model parameters W = tf.Variable([.3], tf.float32) b = tf.Variable([-.3], tf.float32) # Model input and output x = tf.placeholder(tf.float32) linear_model = W * x + b Data Variables & Equations
- 15. Programmatic Example - TensorFlow print(sess.run(linear_model, {x:[1,2,3,4]})) y = tf.placeholder(tf.float32) squared_deltas = tf.square(linear_model - y) loss = tf.reduce_sum(squared_deltas) print(sess.run(loss, {x:[1,2,3,4], y:[0,-1,-2,-3]})) fixW = tf.assign(W, [-1.]) fixb = tf.assign(b, [1.]) sess.run([fixW, fixb]) print(sess.run(loss, {x:[1,2,3,4], y:[0,-1,-2,-3]})) Data & Run
- 16. Get started at scale - AWS ML You can get started today without ANY programming AWS ML works purely based on arbitrary CSV 鍖le as input Does own modelling, training based on the content shared Start predicting in minutes Other options are Lex, Poly & Rekognition by Amazon Vision API, Speech API, Translation API by Google
- 18. Questions & Answers Thank 額看顎.