Zahra Dehghanian

AI_booklet_CE-AUT. Booklet and exam of Artificial Intelligence Master Degree at Amirkabir University of technology.

RCALAD. Implementation of RCALAD Model based on "Anomaly Detection Using Complete Cycle Consistent Generative Adversarial Network" paper *Official Code

Scalable_Sequential_Spectral_Clustering. In this project , I implement SeqSC Algorithm based on "Scalable Sequential Spectral Clustering" paper. Also this repo contains implementation of KNN, Spectral Clustering, SVD, SeqKM, SSVD Algorithms and has got a simple GUI to show the Result of each Algorithm. This project was done under supervision of Dr. Amirmazlaghani

CELF. In this repo. , "Cost Effective Lazy Forward Selection" Algorithm is implemented from scratch in python with only numpy library.

DP-FedAvg. In this repo. I implement Federated Averaging (FedAvg) and Differential Privacy-FedAvg (DP-FedAvg) algorithms for a distributed learning scenario on CIFAR-10 dataset. I also analyze the impact of privacy on the model’s accuracy and utility.

DenseNet_Vs._ResNet. In this Repo. DenseNet and ResNet function are compared. first we use transfer learning to initialize the weights and then fine tune them on outdoor and indoor photo from ImageNet dataset. at last we compare this nets in term of accuracy and precision.

Subgraph_Centrality_Measure. implementation of paper "A Family of Centrality Measures for Graph Data Based on Subgraphs"

stock_market. Jupyter Notebook

persian_compiler. final project for course principle of compiler design ,Dr razazi .

classical-modern-Search-artificial-intelligence. This Repo. contains five problems (Rubik ,SAT,Maze ,8Puzzle,WordTable) solved with nine algorithms (A* ,DFS ,IDDFS ,DLS ,BFS ,UniformCost, bidirectionalSearch, Simulated Annealing, and Hill climbing). This project is done for artificial intelligence course under supervision of Dr.Nikabadi .

Data-Usage-Prediction. In this project we expertise different models like : SCINet, Informer, ARIMA, LSTM on Hamrahe-aval data Usage (MCIRD) Dataset. This was an Multivariant time series task which is proposed for Amirkabir Artificial Intelligence Competitions (AAIC)

Bayes_Classifier_Face_Detection. In this Repo. I implement a baysian classifier to detect face in different picture. it was implemented in different section; The description that each part is implemented for, will follow .

Lazy-Hill-Climbing. in this repo. greedy hill climbing and lazy hill climbing is implemented from scratch with only numpy and scipy library. this project is tested on the facebook101-Princeton dataset.

http_over_udp_network. in this project i implement http over UDP and DNS over TCP . this is a final project for course network Dr.sabaei .

NeuralNetwork_on_FPGA. In this repository I implement a basic Neural Network in VHDL . this project has got 3 phase. in first part all function which is needed in future phase is implemented. in second phase a LSTM cell is implemented and at last phase, a simple neural network is built. This is final project for course FPGA , Dr. Sahebzamani.

NumexInterpreter_ProgramingLanguage. In this repository I implement a small interpreter for Numex language. this program was written with functional language in Racket. this is final project for course Design and Implementation of programming language, Dr. Fallah

Poetry_Generator. A simple RNN-based neural network to create Farsi (Persian) poems, In this Repo. Ferdowsi's Poems (Shahnameh ) is created.

pneumonia_diagnosis. In this Repo. we use a ResNet model with transfer learning to classify pneumonia based on chest X-ray image. this is the first project of course "medical imaging" Dr. Rohban.

NCRP_HLDA. In this project, a simple Nested Chinese Restaurant Process is implemented. this is done by using HLDA for topic modeling with Gibbs Sampler. this project demonstrates the BBC Insight Dataset for evaluation.

Self_Organizing_Map. In this Repo. first a SOFM network is implemented to reduce the dimension of pictures and then a multilayer perceptron is used to classify the faces. this project is done on Yale Face dataset : http://vision.ucsd.edu/content/yale-face-database

classify_text. Jupyter Notebook

LSTM_sequence_prediction. In this Reop. I try LSTM Network with 2 different approaches (padding and splitting sequences )to predict which app will be installed by users.

Hidden-Markov-Model. In this Repository, Algorithms of learning HMM with two approaches is implemented: first, learn parameter when having observations and related state of them with Maximum Likelihood Alg. and the second approach in the condition that just have observations with forward-backward Alg. after constructing HMM Viterbi Alg. is implemented to test a sequence of observation.

Content-based-Recommender-System. In this Repository, A simple content-based book recommendation system is implemented. A representation vector is built with a TF-IDF vector of the Authors name concatenated with the title of the book.

Matrix-Factorization-recommender-System. In this repository, I implement a recommender system using matrix factorization. Here, two types of RS are implemented. First, use the factorized matrix for user and item. and second, rebuild the Adjacency matrix. both approaches are acceptable and implemented in this repo. To factorized the matrix, funk-svd Algorithm is used. you can find his implementation on this link: https://github.com/gbolmier/funk-svd

Multilayer_Perceptron. In this Repo. a Multilayer Perceptron is implemented. Both Regression and Classification task is done on yearpredictionmsd dataset.

Exchange-Email-Scanner. this is my internship project . in this project i try to implement a client user in c# wich sign in into exchange 2016 server and check all inbox attachment of emails and in case it finds bad attachment , this code programmaticly deletes it .

Perceptron_Vs._Adaline. Jupyter Notebook

Naive_Bayes_Classifier_Wine. Naive Bayes Classifier with K-Fold Cross Validation and ROC Curves (one Vs. all technique) on Wine Data-set.

ML-From-Scratch. Machine Learning From Scratch. Bare bones NumPy implementations of machine learning models and algorithms with a focus on accessibility. Aims to cover everything from linear regression to deep learning.

the-gan-zoo. A list of all named GANs!

pytorch_GAN_zoo. A mix of GAN implementations including progressive growing

MultiClass_Logestic_Regression. In this Repository, I implement a multi-class LR with binary LR and one-vs-all technique on Mnist dataset. To test the classifier, I show the accuracy and confusion matrix of train and test set and show 25 random data from test set with true and predicted label.

Maximal_frequent_itemset. In this repo. first, I use FP-growth algorithm to find frequent itemset (from pyfpgrowth library). second, find closed frequent itemsets and finally, find Maximal frequent itemset.

Apriori_SON_spark. In this repo. , I implement A-priori and SON Algorithms to find frequent road in real Traffic dataset using Spark.

LensCraft. Python

TFIDF_Spark. In this project I implement tf-idf with spark on Persian Wikipedia to retrieve percentage of similar articles in specific topic.

Apriori. Here is a simple A-priori Algorithm to find frequent Itemset with size 2 and 3. for join, pruning is implemented. after all, the association rule with max confidence is also reported.

Football_Player_Classification. Python

Gradient_Descent. Python

TF-IDF. In this repo I implement a simple TF-IDF algorithm for dataset of news in persian. This project is done for course Information Retrieval under suprivision of Dr. Momtazi.

ganhacks. starter from "How to Train a GAN?" at NIPS2016

Minimum_Distance_Classifier_OCR. In this Repository. I implement a Minimum Distance classifier to detect Numbers in different picture. it was implemented in different section; The description that each part is implemented for, will follow .

Evolutionary_learning. Python

Linear_Transformation. Python

Intelligent-home-security-threats. All project for the Course of research methodology and presentation , Dr. Safabakhsh . In the context of smart home security threats in the objects layer and ways to deal with it

MinHashing_Spark. In this repo. , I implement Cosine similarity and MinHashing function ( with and / or operator on band ) to find similarity to specific road in real Traffic dataset using PySpark.

dollyTheWarior. game maker