Tim Blog

Himmel oder Hölle

optimizer

Gradient Descent Stochastic Gradient Descent Adaptive Learning Rate Method non-gradient Descent Particle Swarm Optimization a population-based method that defines a set of ‘particles’ that explo...

ml predictor

systematically select the best subset of input features for model training to predict the target variable feature selection Embedded methods: built-in into the model training phase of a classi...

Geocoding

map geo data on map get geo data GeoPy from geopy.geocoders import Nominatim #openstreet map geolocator = Nominatim(user_agent="example app") #google map geolocator = GoogleV3(api_key=AUTH_KEY) geo...

federated learning

collaboratively learn a shared prediction model while keeping all the training data on the local device

clustering

tSNE tSNE is the t-distributed stochastic neighborhood embedding PCA works by preserving the variability of the data; tSENE works by non-linear probility PCA focusses on global structure rathe...

puppeteer

A Node library which provides a high-level API to control headless Chrome or Chromium over the DevTools Protocol setup environment #npm install puppeteer usage build API #an example api to co...

curriculum learning

it lets the model learn from easy samples first, and gradually advance to complex samples and knowledge.

point cloud processing

Object Detection

facial emotion recognization

Facial Detection APIs that Recognize Mood Emotient Affectiva EmoVu

meeting summary

There are two approaches for text summarization: abstractive summarization generate new sentences from the original text extractive summarization identify the important sentences...