ML - The way the world works - analyzing how things work

Latest Episodes:

Fully self driving cars will become the best option for navigation Oct 05, 2019

Show notes

Fsd vision system will need to rely on radar less and image processing more. The Tesla vehicle should not curb hug; it should avoid brake tapping if it thinks and object might enter it path but there was no intersection possibility ; it should resume target speed immediately after a block car changes lanes; it should not wander in the drive lane; it should accurately identify cars in the rear quadrant. Tesla should collect data from all cars and store it on the cloud for situation training automatically . Image processing should be watching far ahead to recognize when highway traffic has abruptly stop. It should watch other cars hitting brake lights to see a mass car slowdown occurring.

Acne Hansen talks about Dynamic IQ Sep 25, 2019

Show notes

Taking care of money

Biren Gandhi talks about the $19 trillion internet of things market Sep 24, 2019

Show notes

about IoT/Voice.

Deep learning and reinforced learning by Asheeh Kumar Sep 21, 2019

Show notes

Building deep Learning-based classifiers for the task of 'Emotion Classification' using Pytorch classifiers for Emotion. • Used CNN Lstm VGG16 Pre-trained model and Attention form Paper "Attention is all you need". • Multi-modal classifiers for disaster tweet classification using tweet text and image. • Experiments for Multi-lingual Multimodality for emotion classification and predicting Intensity score. • Evaluating all results using precision-recall and F1-score cosine-similarity and Pearsons correlation coefficient Artificially generate realistic labelled dataset | SWAAYATT ROBOTS 16th May 2018 - 10th July 2018 • Solved the problem of Night Vision using Multimodel Unsupervised Image to Image translation. • Implemented GANs and Autoencoder for style transfer to artificially generate realistic labelled datasets. • Generated Inverse Perspective Mapping software(C++) using USB cameras. Projects Road Accident analysis and Safety measures. | IIT Roorkee Feb-2019 March-2019 • Analyzing accident data and coming up with important factors in an accident by visualizing of road accident data from 2003-2016. • Building a Faster R-CNN model for Traffic sign Detection and classification using pytorch and Use this model to give warning on not following signs. Text to Narrate | IIT Roorkee Artificial Intelligence and Electronics Society Dec 2017 - Jan 2018 • Implemented bi-LSTM layer, Conditional Random Field (CRF) Named Entity Recognition. • Render related images to the user so as to visualize the story. • The final phase of the project was to combine the two sub-parts to provide a seamless experience for the end-user. • We have made the web-app with the help of Django.

Journey to Mars with Ben Taylor Sep 17, 2019

Show notes

Is it possible to colonize Mars . What are the challenges to terraforming mars atmosphere? What are the financial incentives to creating a journey to mars robot or human

Emergence : ants , big cities , and brains are the pattern ai agents use to learn today Sep 16, 2019

Show notes

Ants and terminates make up 30 percent of the Amazonian rain forest biomass. There are nearly 10,000 species of ant. Ants lack any advanced forebrains. “While there’s no single key to the success of the social insects, the collective intelligence of the colony system certainly played an essential role.” “Call it swarm logic”. A harvest ant will not only ascertain the shortest route to a food source, it will also prioritize food sources, based on their distance and ease of access. In response to changing external conditions, worker ants switch from nest-building to foraging to raising ant pupae. None of the individual ants are in charge. Deborah Gordon became interested and said, “I was interested in system where individuals who are unable to assess the global situation still work together in a coordinated way, and they manage to do it using only local information.” Harvest ants constant adjust the number of ants actively foraging for food, based on a number of variables: the overall colony size, the amount of food stored in the nest, the amount of food available in the surrounding area, even the vicinity of other colonies in the near vicinity. Ants secrete a finites number of chemicals and these chemical signals are used to communicate with other ants. “A foraging ant might expect to meet three other forages per minute-if she encounters more than three, she might follow a rule that has her return to the nest. Because larger, older colonies produce more foragers, ants may behave differently in larger colonies because they are more like to encounter other ants. Base on pheromone signals and frequency, the ant can adjust its behavior.

Abstraction is the only way to prove a theory and philosophy can only criticize not prove Sep 16, 2019

Show notes

Proof is in abstraction. So philosophy is either self-evident or it is not philosophy. Philosophy can only argue that the existence of God is self-evident. Philosophy does not have the capacity to prove the nonexistence of God because it cannot abstract. Objectivity at best can only criticize those who believe in God and this criticism can be classified as "lacking understanding".

Logistic regression Sep 14, 2019

Show notes

Gradient convergence

Support vector machine and linear discriminant Sep 10, 2019

Show notes

The support vector machine is a linear discriminant classifer. SVM separates the class entities by a linear discriminant line. The fatter the margin between the discriminant and the class entity and the better the lower the penalty for error for class entity on the wrong side then the better the prediction.

Restrictive Boltzmann Machines are better than back error propagating neural network Sep 04, 2019

Show notes

The rbm is unsupervised learning system . Multiple rbm can be combined to reduce error.

Our TOPPODCAST Picks

Follow Us

Stay Connected

ML – The way the world works – analyzing how things work

Related Podcasts

Links

Stay Connected