Team Projects

Fake news is ever more powerful when it comes to riling up the populus in hysteria. With this issue in mind, in this project, we sought to produce a program based on machine learning models capable of assessing online articles' credibility. The project aimed to assess whether machine learning models and NLP algorithms could be utilized to produce software in which the user will enter a link to an article, and the program will output the article's credibility score with regards to its sources, text, and how it may compare to other credible articles, providing a relatively intuitive approach to this issue. After developing our model which used multiple different corpora to analyze different aspects of an article and then compares these factors to develop a final score, we found our results to be reasonably well, with a mean final percent error of 9.72% within our acceptable range of 10%. At the same time, our results revealed to us which factors we were using that needed revision or were not meaningfully impacting our final result. As a consensus of our results, we found that our model for assessing website credibility is a viable solution for the problem of misinformation spread throughout the internet. Although our model is not perfect yet, it performs decently well, and with more data, development, and training it can significantly improve in accuracy and feature extraction.

Modern technology has led to misinformation on social media, especially with regard to the coronavirus pandemic, which has become a pertinent issue, and it can be challenging to find credible, informative sources on the platform. A tool must allow individuals to assess the credibility of everyday posts and sources to aid in getting credible information.

This project aims to assess whether Machine Learning models and Natural Language Processing (NLP) algorithms can approximate an article’s credibility score by comparing its sources, text, and information to other credible articles. The objective is for the program to output valid credibility scores for over 75% of articles, utilizing corpuses such as the Web Content Credibility Corpus for testing our program.

We determine the credibility of articles using natural language processing and matching learning evaluation. Specifically, we extract knowledge graphs from unstructured text using Transformers and the "Attention is all you need" framework. Using this, we are able to construct a knowledge graph of known information, as well as a knowledge graph of a specific article, using these knowledge graphs, we are able to compare information to evaluate the credibility of an article.

What is Climap?

Climap is an online interactive visualizer to predict how climate change will impact the world. Current trends indicate a rise in both global temperatures and pollution, specifically greenhouse gases within the next couple of decades.

User Guide

You can try out our map under the explore section of our website. Because much of this is thought to be caused by human activities, we have overlayed human population density on top of the map. Drag the slider at the bottom of the screen to view predictions of future greenhouse gas levels in varying countries around the world.

You can try it out on our website! https://pollution.theglobaltech.org/

See this video about our project! https://www.youtube.com/watch?v=5diupKrQ3Uc

Methodologies

In order to predict the effect of climate change on various countries and regions around the world, we have developed a model using regression to predict future average carbon dioxide levels. We plan add more features in the future, and to scale this to specific cities as opposed to just countries. Please see the references section to view the databases and frameworks used for this project.

Credits and References

This project was developed during the Augmented Hacks Hackathon. https://www.augmentedhacks.org/

For a further list of references, see here. https://github.com/alantian2018/Augmented_Hacks/blob/main/References.md

Inspiration

I'm sure everyone has heard of global warming, climate change, and the need to reduce greenhouse gases in the atmosphere. While traditional methods, such as rallying others through social media or hosting clean-ups are essential, we wanted to apply computing power to be able to better address this issue.

What it does

The user will upload an image to our website. This image is fed through a neural network and different pieces of plastic will be identified.

How we built it

We used Django as a framework. All web pages were written from scratch. The model utilized Google's Inceptionv3 image classification model. The library used was Tensorflow.

Challenges we ran into

The biggest challenge was training a model that worked. We tried many different models, such as resnet18 (75% accuracy), mobile-net (60% accuracy), and even tried writing a custom network. We tried implementing them in both Pytorch and Tensorflow. Unfortunately, all failed to achieve an acceptable accuracy. Even our final model, Inception, only achieved around an 80% accuracy.

Another challenge we faced was connecting our frontend and backend. We had a hard time passing the image from the upload interface to our model input.

Accomplishments that we're proud of

While we knew that training the model would be the most difficult task, we were proud that we were able to write almost all the code for the website within the first day.

What we learned

We learned a lot about web development during this Hackathon. Many of us have had prior experience with machine learning and training models, but this Hackathon was a great way for us to gain experience implementing a website using HTML/CSS.

What's next for TheLifeCycle

Our goal is to create a model capable of achieving multiclass classification, which would be capable of telling the user what type of trash they have uploaded (ie: paper, plastic, cardboard, etc) instead of telling the user if it is only recyclable/organic. In addition, we would like to increase our current model's classification accuracy.

Hackathon Page (Devpost)
https://devpost.com/software/thelifecycle#updates