This repository contains the daily updates made during the 60daysofudacity challenge. This was an initiative from the Secure and Private AI Challenge Scholarship program.
Began with the 60daysofudacity Challenge, completed Lesson 1-4
- L-1: Introduction
- L-2: Deep learning with Pytorch
- Introduction to Pytorch
- Introduction to Neural Networks
- Training and Loading the data
- Inference and validation
- Transfer learning
- Project: Train a pretrained models to classify the cat and dog images
- MNIST and FNIST dataset
- L-3: Introduction to Differential Privacy
- Anonymous and canonical data
- Project - Generate Parallel Databases
- L-4: Evaluation of Privacy Function
- Project - Evaluating the Privacy of a Function
completed lesson 5.1 to 5.5
- Introduction to Local and Global Differential privacy
- Project: Calculate L1 Sensitivity For Threshold
- Project: Perform a Differencing Attack on Row 10
- Project: Local Differential Privacy
- References:
- https://medium.com/snips-ai/differential-privacy-for-the-rest-of-us-665e053cec17
- https://desfontain.es/privacy/local-global-differential-privacy.html
- https://stats.stackexchange.com/questions/370591/what-are-global-sensitivity-and-local-sensitivity-in-differential-privacy
- http://jmlr.org/papers/v17/15-135.html
- Meetup with #sg_wrldwde-women-cdrs : discussed Dogs vs Cats Kaggle dataset in PyTorch.
- Looked into DATA Augmentation
- References:
- https://stackoverflow.com/questions/51677788/data-augmentation-in-pytorch
- https://medium.com/ymedialabs-innovation/data-augmentation-techniques-in-cnn-using-tensorflow-371ae43d5be9
- https://colab.research.google.com/drive/109vu3F1LTzD1gdVV6cho9fKGx7lzbFll
- https://pythonzeal.wordpress.com/2018/07/06/data-augmentation-with-pytorch/
- https://www.datascience.com/blog/transfer-learning-in-pytorch-part-one
- https://medium.com/@thimblot/data-augmentation-boost-your-image-dataset-with-few-lines-of-python-155c2dc1baec
- Participated in Virtual Hackathon on #sg_hackathon_orgnizrs i.e., SPAIC HACKATHONERS Hackathon Blossom (Flower Classification)
- Link: https://www.kaggle.com/spaics/hackathon-blossom-flower-classification
- Continued working on the Hackathon Blossom (Flower Classification), successfully trained the model and validated it.
- working on testing the trained model with the given test set!
- Attended webinar of Robert Wagner
- Revised all the course work completed up till today!
- In the hackathon code predicted and displayed the top 5 from the list but have to change it to display the predicton for all the test set!
- Continued working on hackathon!
- Looked into cats and dogs challenge as decided by #sg_wrldwde-women-cdrs
- References:
- https://heartbeat.fritz.ai/brilliant-beginners-guide-to-model-deployment-133e158f6717
- https://medium.com/dair-ai/a-simple-neural-network-from-scratch-with-pytorch-and-google-colab-c7f3830618e0
- https://www.analyticsvidhya.com/blog/2019/01/guide-pytorch-neural-networks-case-studies/ , https://skymind.ai/wiki/neural-network
- https://adventuresinmachinelearning.com/convolutional-neural-networks-tutorial-in-pytorch/
- completed the Hackathon, had to resend the predictions for all test sets!
- looked into lesson 5 - Introduction to Local and Global Differential privacy
- Updating my github with all the course exercises and ai projects
- practice lesson 5.7 and 5.8
- Project - Project: Varying Amounts of Noise
- Project: Create a Differentially Private Query
- looked into posts to deploy deep learning projects like like flower classifier to be deployed as web app or mobile app
- References:
- https://heartbeat.fritz.ai/brilliant-beginners-guide-to-model-deployment-133e158f6717
- https://medium.com/@daj/creating-an-image-classifier-on-android-using-tensorflow-part-1-513d9c10fa6a
- https://medium.com/datadriveninvestor/how-to-build-a-machine-learning-app-choosing-the-best-image-recognition-api-297d4d97c84c
- https://reshamas.github.io/deploying-deep-learning-models-on-web-and-mobile/
- https://heartbeat.fritz.ai/easily-build-image-classification-models-using-just-your-smartphone-walkthrough-guide-43bf211181b0
- https://blog.floydhub.com/build-image-classification-app-with-fastai/
- working on the task 1 as per #sg_real_world_ai_proj
- aim is to complete this Generative Adversarial Networks for beginners tutorial https://www.oreilly.com/learning/generative-adversarial-networks-for-beginners, before tomorrows meetup!
- Updating my Github account from the very beginning of this challenge!
- References:
- as per study group #sg_real_world_ai_proj learning about GAN and its implementations
- References:
- https://www.oreilly.com/learning/generative-adversarial-networks-for-beginners
- https://www.analyticsvidhya.com/blog/2017/06/introductory-generative-adversarial-networks-gans/
- https://github.com/nashory/gans-awesome-applications
- https://towardsdatascience.com/machine-learning-for-cybersecurity-101-7822b802790b
- https://towardsdatascience.com/generative-adversarial-networks-gans-for-beginners-82f26753335e
- https://sites.google.com/view/cvpr2018tutorialongans/
- https://sites.google.com/view/cvpr2018tutorialongans/
- https://machinelearningmastery.com/resources-for-getting-started-with-generative-adversarial-networks/
- Learning more about GAN
- References:
- Found this awesome video explanation by Siraj Raval https://www.youtube.com/watch?v=yz6dNf7X7SA
- https://medium.com/@devnag/generative-adversarial-networks-gans-in-50-lines-of-code-pytorch-e81b79659e3f
- https://medium.com/@awjuliani/generative-adversarial-networks-explained-with-a-classic-spongebob-squarepants-episode-54deab2fce39
- Another awesome video help me understand CNN better by Siraj RAval Convolutional Neural Networks https://www.youtube.com/watch?v=FTr3n7uBIuE
- And on NN https://www.youtube.com/watch?v=h3l4qz76JhQ&list=PL2-dafEMk2A5BoX3KyKu6ti5_Pytp91sk
- checked this repo https://github.com/llSourcell/Learn_Deep_Learning_in_6_Weeks
- did revision!
- As per study group #sg_real_world_ai_proj learning about GAN and its implementations, looked into following good resources:
- References:
- https://www.oreilly.com/learning/generative-adversarial-networks-for-beginners
- https://www.analyticsvidhya.com/blog/2017/06/introductory-generative-adversarial-networks-gans/
- https://github.com/nashory/gans-awesome-applications
- https://towardsdatascience.com/machine-learning-for-cybersecurity-101-7822b802790b
- https://towardsdatascience.com/generative-adversarial-networks-gans-for-beginners-82f26753335e
- https://sites.google.com/view/cvpr2018tutorialongans/
- https://sites.google.com/view/cvpr2018tutorialongans/
- https://machinelearningmastery.com/resources-for-getting-started-with-generative-adversarial-networks/
According to the challenge rules: Break of not more than 1 day within two weeks of time is allowes! And because of my caollege final project and exams the strike was broken! Took 5 days break! So had to restart again!
- missed out for about a week coz of college work and starting the challenge again... Completed Lesson - 5
- Project - PATE Analysis
- Completed lesson 6 - Differential Privacy for Deep learning!
- Final project on Differential Privacy - Train a DP model using this PATE method on the MNIST dataset
- Began working on aws deepracer challenge!
- Started lesson 7 on Federated learning looked into the following blogposts for better understanding
- Project: Playing with Remote Tensors using Pysft
- Advanced Remote Execution Tools
- References:
- https://medium.com/syncedreview/federated-learning-the-future-of-distributed-machine-learning-eec95242d897
- https://hackernoon.com/a-beginners-guide-to-federated-learning-b29e29ba65cf
- https://hackernoon.com/a-beginners-guide-to-federated-learning-b29e29ba65cf
- https://medium.com/datadriveninvestor/an-overview-of-federated-learning-8a1a62b0600d
- https://www.kdnuggets.com/2019/06/pysyft-emergence-deep-learning.html
- https://towardsdatascience.com/private-ai-federated-learning-with-pysyft-and-pytorch-954a9e4a4d4e
- Completed lesson 8 - Securing Federated Learning
- Project: Federated Learning with a Trusted Aggregator
- Project: Build Methods for Encrypt, Decrypt, and Add
- Final Project: Federated Learning with Encrypted Gradient Aggregation
- Started lesson 9 - Encrypted Deep Learning
- Reviewing Additive Secret Sharing
- Encrypted Subtraction and Public/Scalar Multiplication
- Encrypted Computation in PySyft
- Completed lesson 9 - Encrypted Deep Learning
- Project: Build an Encrypted Database
- Encrypted Deep Learning in Keras
- Project: Private Prediction using Syft Keras - Serving (Client)
- Completed the course!
- Mostly revised some topics from the course!
- Searched for project ideas to work on, still not sure what i want to do for the project so am looking into online papers and other projects on github based on GAN or other DL techniques.
- worked on aws deepracer
- Reading few more post to understand the concepts better!
- References:
- https://medium.com/apache-mxnet/epsilon-differential-privacy-for-machine-learning-using-mxnet-a4270fe3865e
- https://www.kdnuggets.com/2015/01/differential-privacy-data-mining-compatible.html
- https://www.researchgate.net/publication/269997816_Differential_Privacy_and_Machine_Learning_a_Survey_and_Review
- https://www.cylance.com/en-us/resources/knowledge-center/ai-and-ml-for-security.html
- https://towardsdatascience.com/machine-learning-for-cybersecurity-101-7822b802790b
- Showcase Project related anouncements were made!
- Searched the web for better ideas for showcase project
- looked deeper in GAN and came across google lens concepts
- looked into past hackathon challenges to practice on and build solid dl foundation in the coming days.
- Working on some key concepts to polish my knowledge of dl from the course again!
- Worked on first hackthon challenge
- working on showcase project , referred to the following reseources
- References:
- https://firebase.google.com/docs/ml-kit
- https://codelabs.developers.google.com/codelabs/mlkit-android/#0
- https://firebase.googleblog.com/2018/05/introducing-ml-kit-for-firebase.html
- https://medium.com/coding-blocks/creating-a-qr-code-reader-using-firebase-mlkit-60bb882f95f9
- https://becominghuman.ai/generative-adversarial-networks-for-text-generation-part-1-2b886c8cab10
- Working on showcase project, looking for ways to deploy my pytorch code in mobile app!
- looked into the following resources so far but will try it out tomorrow, there are so many kinda confused to what i can use and what i can't for the project!
- Setting up my write-ups and walk through for the course, will put it soon on github!
- References:
- https://developers.facebook.com/blog/post/2018/05/02/announcing-pytorch-1.0-for-research-production/ , https://github.com/onnx/tutorials
- https://github.com/cedrickchee/pytorch-android
- https://pytorch.org/tutorials/advanced/cpp_export.html
- https://firebase.google.com/docs/ml-kit/use-custom-models
- https://heartbeat.fritz.ai/deploying-pytorch-and-keras-models-to-android-with-tensorflow-mobile-a16a1fb83f2
- https://pytorch.org/tutorials/advanced/super_resolution_with_caffe2.html
- working on showcase project idea, worked on content for my git repo with walk through for this course, will upload pretty soon!
- came across this awesome blog post of Adam Geitgey on medium called ***Machine learning is fun! ***
- References:
- https://medium.com/@ageitgey/machine-learning-is-fun-80ea3ec3c471
- https://medium.com/@ageitgey/natural-language-processing-is-fun-part-3-explaining-model-predictions-486d8616813c
- https://medium.com/@ageitgey/how-to-break-a-captcha-system-in-15-minutes-with-machine-learning-dbebb035a710
- https://medium.com/@ageitgey/quick-tip-the-easiest-way-to-grab-data-out-of-a-web-page-in-python-7153cecfca58
- https://www.hackevolve.com/building-a-reverse-image-search-engine/
- working on the project , looked in for some research papers and articles for the project.
- References:
- working on showcase project idea! Based on Tensorflow and Pytorch
- working on project found some great resources
- References:
- https://github.com/CSAILVision/semantic-segmentation-pytorch
- https://www.analyticsvidhya.com/blog/2019/08/introduction-slimyolov3-real-time-object-detection/
- https://www.analyticsvidhya.com/blog/2018/10/a-step-by-step-introduction-to-the-basic-object-detection-algorithms-part-1/?utm_source=blog&utm_medium=7-innovative-machine-learning-github-projects-in-python
- https://blog.paperspace.com/how-to-implement-a-yolo-object-detector-in-pytorch/
- http://www.covert.io/security-data-science-learning-resources/
- working on the project , looked up on few articles on ***object detection ***
- References:
- worked on showcase project and Onto submitting my showcase project!
- Submitted two projects!
- Real Time Object Detection with Pytorch
- Guide
- Just found four project folders in our course material, worked on that !
- project-bikesharing
- project-dog-classification
- project-face-generation
- project-tv-script-generation
- Looked through few papers and post on Scene text detection
- References:
- https://www.learnopencv.com/deep-learning-based-text-detection-using-opencv-c-python/
- https://medium.com/syncedreview/stn-ocr-a-single-neural-network-for-text-detection-and-text-recognition-220debe6ded4
- https://engineering.fb.com/ai-research/rosetta-understanding-text-in-images-and-videos-with-machine-learning/
- https://paperswithcode.com/paper/east-an-efficient-and-accurate-scene-text.
- Wrote an introductory series on tensors on medium - Introduction to "Tensors" - 1
- revised the course material
- Wrote two more posts of introductory series on tensors on medium!
- Introduction to "Tensors" - 2 (Using Pytorch)
- Introduction to "Tensors" - 3 (Using Pytorch)
- Looked somemore into YOLO, for my project on object detection!
- Wrote the last post on tensor series - Introduction to "Tensors" - 4 (Using Pytorch)
- Link: https://medium.com/@shaistha24/introduction-to-tensors-4-using-pytorch-8c84f035268b
- Worked on the object detection project by Yolo!
- Revised course on differential privacy!
- Looked up a bit more on yolo for object detection, adding (in process of content creation for display, pls support and give a clap if you like it later when i add it) my showcase projects ( Guide and Real Time Object Detection Using Pytorch) details to the #project-showcase-challenge
- looked more into object detection referred to the following
- References:
- https://medium.com/coinmonks/detecting-custom-objects-in-images-video-using-yolo-with-darkflow-1ff119fa002f
- https://cv-tricks.com/object-detection/faster-r-cnn-yolo-ssd/
- https://hasgeek.com/fifthelephant/2019/proposals/machine-learning-security-the-data-scientists-guid-uxPNxWYFypavc6xnNHdvPW , and a good read on federated learning on hackernews
- https://news.ycombinator.com/item?id=19944510
- Wrote another blog post on Basic Concepts/ You Should Know Before Starting with the “Neural Networks” (NN) — 1
- Had a meetup with @Helena Barmer @Jess @Munira Omar @Seeratpal K. Jaura @Temitope Oladokun @Astha Adhikari @Suparna S Nair @Labiba regarding udacity thank you vdo creation!
- References:
- Wrote the second article on medium on ***Basic Concepts You Should Know Before Starting with the “Neural Networks” (NN) — 2 ***
- Began Updating my github for #60daysofudacity challenge
- worked on my Showcase project Guide!
-
Made it to the Honorable Mentions list Ranked 14th for the Showcase Project with - Real Time Object Detection with Pytorch
- Link: https://github.com/aksht94/UdacityOpenSource/tree/master/Shaistha/real_time_object_detection_using_pytorch
- All the project submitted for the challenge can be found here
- Link: https://github.com/aksht94/UdacityOpenSource/tree/master/Shaistha/real_time_object_detection_using_pytorch
-
Updated my github repository for 60daysofudacity challenge!
-
Submitted the form with 60daysofudacity challenge!
-
Updating my github repo for Deep-learning-with-Pytorch
This section contains the projects developed during the Challenge.
| Project Name | Description |
|---|---|
| Classifying MNIST | Neural Network project for MNIST dataset. |
| Cat and dog image Classifier | Learnt how to use pre-trained networks to solved challenging problems in computer vision. Specifically, you'll use networks trained on ImageNet available from torchvision. |
| Classifying Fashion-MNIST | Build and train a neural network using FMNIST dataset. |
| MNIST_GAN | Building a generative adversarial network (GAN) trained on the MNIST dataset. |
| Multi-Layer Perceptron, MNIST | Train an MLP to classify images from the MNIST database hand-written digit database. |
| Flower power | Using VGGNet to classify images of flowers. |
| Predicting Student Admissions with Neural Networks | Neural network implementation. |
| CNN-CIFAR-10 database Classifier | CNN to classify images from the CIFAR-10 database. |
| Sentiment_Classification_Projects | Practicing Sentiment Classification & How To "Frame Problems" for a Neural Network |
| by Andrew Trask | |
| Sentiment Analysis_RNN | implement a recurrent neural network that performs sentiment analysis. |
| Skip-gram Word2Vec | Implement the Word2Vec algorithm using the skip-gram architecture and learn about embedding words for use in natural language processing. |
| Character-Level LSTM | Constructing a character-level LSTM with PyTorch. The network will train character by character on some text, then generate new text character by character. |
| CycleGAN-Image-to-Image Translation | Define and train a CycleGAN to read in an image from a set 𝑋.X and transform it so that it looks as if it belongs in set 𝑌. Y Specifically, looking at a set of images of Yosemite national park taken either during the summer of winter. The seasons are our two domains! |
| DCGAN | Training DCGAN on the Street View House Numbers (SVHN) dataset. These are color images of house numbers collected from Google street view. SVHN images are in color and much more variable than MNIST. |
| Differential Privacy in the context of a database query-projects | learn how to know whether a database query over such a small database is differentially private or not - and more importantly - what techniques are at our disposal to ensure various levels of privacy. |
| Create a Differentially Private Query | create a query function which sums over the database and adds just the right amount of noise such that it satisfies an epsilon constraint. |
| Differentially Private Final Project | Train a DP model using this PATE method on the MNIST dataset |
| Federated Learning | Techniques for training Deep Learning models on data to which you do not have access. Remote Access. |
| Federated Learning with a Trusted Aggregator | trusted aggregator : a neutral 3rd party who has a machine that we can trust to not look at the gradient when performing the aggregation.Rather than depennding on the db owner who might be inlined to other data partners , we can choose a 3rd neutral party ...anyone in the planet to larger pool to find the trust worthy person. Send all their grad to neutarl 3rd party |
| Securing Federated Learning | Implements final project for Lesson 8, where secure aggregation is added to Federated Learning. |
| Encrypted Deep Learning | Implements final project for Lesson 9, where encryption is added to Federated Learning. |
| Encrypted Deep Learning in Keras | Same using Keras |
| Private Prediction using Syft Keras - Serving (Client) | To the above keras project, securing it with Syft Keras, and make predictions |
| Predicting Bike Sharing | Neural network and use it to predict daily bike rental ridership. |
| Dog Identification-CNN | Algorithm for a Dog Identification App |
| Face Generation | Train a DCGAN on a dataset of faces.Goal is to get a generator network to generate new images of faces that look as realistic as possible! |
| TV Script Generation | Generate Seinfeld TV scripts using RNNs.The Neural Network will generate a new ,"fake" TV script, based on patterns it recognizes in this training data. |
| Showcase-Project 1 - Real Time Object Detection with Python | All the info can be found here: https://github.com/aksht94/UdacityOpenSource/tree/master/Shaistha/real_time_object_detection_using_pytorch |
| Showcase-Project 2 - Guide | All the info can be found here: https://github.com/aksht94/UdacityOpenSource/tree/master/Shaistha/guide |