saurabhmathur96

To Trust Or Not To Trust A Classifier. A measure of uncertainty for any trained (possibly black-box) classifier which is more effective than the classifier's own implied confidence (e.g. softmax probability for a neural network).

Language:Jupyter NotebookApache-2.0173 13 3

natural-gradients

Collection of algorithms for approximating Fisher Information Matrix for Natural Gradient (and second order method in general)

Language:PythonBSD-3-Clause137 4 4

Better_LSTM_PyTorch

An LSTM in PyTorch with best practices (weight dropout, forget bias, etc.) built-in. Fully compatible with PyTorch LSTM.

Language:PythonMIT132 4 6

asnr

Animal Social Networks Repository

Language:PythonGPL-3.060 7 3

limitations-empirical-fisher

Limitations of the Empirical Fisher Approximation

Language:Python44 40

AdaGCN

reproduce ICLR2021 paper "AdaGCN: Adaboosting Graph Convolutional Networks into Deep Models"

Language:Python18 10

deep-learning-uncertainty

My master's thesis project

Language:TeX9 30

myrtlespeech

Speech recognition

Language:PythonNOASSERTION8 5 5

PyIBP

An Indian Buffet Process implementation in Python.

Language:Python7 20

OrgProbe

A linux command-line probe that checks whether a website is blocked by filters and reports results back to the censorship monitoring project middleware.

Language:PythonGPL-3.06 12 5

carl-lib

Causal Inference on Relational Data

Language:Racket4 5 14

relational-datasets

Python package for fetching and using srlearn-compatible relational datasets.

Language:PythonApache-2.04 1 10

autopgm

Automatically train a global Bayesian Network from multiple heterogeneous data sources

Language:PythonMIT3 30

watchdog

Censorship detection framework

2 2 2

CoffeeMachine

A coffee machine which can serve beverages in parallel. Submitted for Dunzo

Language:Java1 20