Artificial Intelligence

Get the most out of your training, scoring, algorithms and frameworks on Intel® architecture for Deep Learning and Artificial Intelligence.

Activity Feed

Medium mhfnw 1c

Niladri S. created project XFakeNewsChecker

Medium 0c74cf56 dd8a 453b 9584 af258c4989a0


This project was made by me and a couple of my friends at IET Hackathon 2018 in Chennai. We won the 2nd prize.

Problem Statement Often sensational news is created and spread through social media to achieve intended end. On the other hand, it may also involve narration of a true fact however being deliberately exaggerated. This may also affect the importance of serious news media. The problem is to identify the authenticity of the news and online content. Equally important problem is to identify the bots involved in spreading false news. Approach The problem can be broken down into 3 statements :-

Use NLP to check the authenticity of a news article.

If the user has a query about the authenticity of a search query then he/she can directly search on our platform and using our custom algorithm we output a confidence score.

Check the authenticity of a news source.

Articles can be analyzed by feeding them to a machine learning model (Passive Aggressive Classifier) which predicts the genuinity of the content after it's trained through predefined datasets of classified real vs fake news. Search terms can be analyzed by doing a Google search (first 100 entries) and and ensuring if the news corresponding to the keywords have been covered by reliable news sources and aggregators. For every search term covered by a reliable news source it recieves a score of +1, while we heavily penalize fake sources. If multiple fake sources cover the news then we penalize the truth score even harder. We also look for keywords like 'hoax', 'fake', etc in the payload content. An URL (news source) can be analyzed if it's authentic by checking it in our database of true news provider and false news provider. Application When the fake-news detector is hosted locally / on a cloud platform, it can predict fake news with reasonable accuracy. Since edge cases of fake-news detection are controversial, the tool outputs a probability percentage instead of a rigid label.

Default user avatar

高 鹏. created project Smile Baby

Medium c85c3d39 5c33 4299 8745 9968744b57c1

Smile Baby

Baby crying real-time monitoring device. The hardware is responsible for monitoring the baby cries when the parents are not around the baby, analyzing the baby cries by the server, and feeding the obtained crying results to the baby guardian's cell phone to facilitate the newborn's parents to take care of the babies.

Default user avatar

chunxu z. created project iMsg

Medium 8719d78d 90c1 4592 951f 9384d293de0e


Nowadays, we college students join a lot of QQ chat groups to receive useful information such as announcements made by the school. But to our frustration, people sometimes send garbage messages in these groups, making it harder for us to find information that's actually useful, and it can also cause us to miss many opportunities, defeating the intended purposes of these groups. Now we are developing a software that can change this situation. The software first retrieves QQ chat logs to analyze and extract information from them, and then uploads records of important information to the cloud, enabling users to view them across their devices and remind them of their to-do items. For this project, we utilizes Intel's PyTorch to create neural networks in order to analyze messages, and we use Microsoft Azure to host our web services that allow users to view the records anytime.

Medium pp

Avideep M. updated status

Medium pp

Avideep Mukherjee

I am currently working on kNN classifiers. I want to know a simple case where weighted kNN classifier outperforms kNN classifier. Any help would be appreciated.

Default user avatar

AMIRSINA T. created project Lip Reading - Cross Audio-Visual Recognition using 3D Convolutional Neural Networks

Medium 1f56e0df 5afe 42cb 8546 bd16cc283223

Lip Reading - Cross Audio-Visual Recognition using 3D Convolutional Neural Networks

Audio-visual recognition (AVR) has been considered as a solution for speech recognition tasks when the audio is corrupted, as well as a visual recognition method used for speaker verification in multispeaker scenarios. The approach of AVR systems is to leverage the extracted information from one modality to improve the recognition ability of the other modality by complementing the missing information. The essential problem is to find the correspondence between the audio and visual streams, which is the goal of this paper. We propose the use of a coupled 3D convolutional neural network (3D CNN) architecture that can map both modalities into a representation space to evaluate the correspondence of audio-visual streams using the learned multimodal features. The proposed architecture will incorporate both spatial and temporal information jointly to effectively find the correlation between temporal information for different modalities. By using a relatively small network architecture and much smaller dataset for training, our proposed method surpasses the performance of the existing similar methods for audio-visual matching, which use 3D CNNs for feature representation. We also demonstrate that an effective pair selection method can significantly increase the performance. The proposed method achieves relative improvements over 20% on the equal error rate and over 7% on the average precision in comparison to the state-of-the-art method.

Medium yoshikawa.chinatsu.full.620099

Yuanzheng C. created project ColorEvangelist

Medium e232d9bd aae4 49c3 a274 a7c0f6a1d912


Sketch/Line Art colorization is a time consuming process, The automation of the process requires not just simple boundary detection but also semantical feature identification, user interaction and shading(which is not a problem for grey image colorization). We propose an deep end-to-end trainable colorization model that meanwhile small in size and has a almost-real-time performance. It identifies semantic features from the sketch and inpaint it with/without user interaction with realistic shading.

See More

Projects (535)

See All

See More

Thumb 8d3f9093 8bc6 40ff 8255 f5421e9631d6
  • Collaborators 0
  • Followers 0

Triton UAS

Default user avatar Marco Flowers

Created: 03/03/2017

Triton UAS is an engineering student organization at UC San Diego that develops deep learning sys...

See More