This project requires knowledge of the following subjects:
Project objectives:
The goal of the project is to build a more contextual search engine for the Arabic language. Usually search engines for the Arabic language relies on the existence of the search term in a document to retrieve the document. The goal of this project is to also retrieve documents that contains some context of the search term and score it. To get this context the persisted document and the search term has to go through a processing pipeline to normalize the text. The processing pipeline has to do some text tokenization, lemmatization, remove stop words, and capture similarity between words.
Requirements:
The end product is preferred to be some kind of an extension on Spacy or Gensim or NLTK that provides the same functionalities they currently have for the Arabic language. That would enable to apply advanced NLP tasks on an Arabic text, mainly Tokenization, Lemmatization, Stop words removal, words similarities, stemming etc….
Project deliverables:
==== Project 2: Smart Bottle ====
This project requires knowledge of the following subjects:
Project objectives:
The goal of the project is to help people regulate their fluid intake whether it is water, juice, coffee or any other drink.
So we need to create a smart bottle and a companion app that indicate how much fluid the user drank over a certain period of time (day, week, etc..).
Requirements:
Project deliverables:
Constraints:
Bonus points:
Should you need additional information, please let me know.
Murex has a considerable internal community of developers with different skill sets and levels. We need to have a continuous overview of which skills are most common and which are sorely lacking, so that we can better predict our staffing needs on various projects. We want to implement a software solution that automates this journey, prompting collaborators to assess their competencies on skills of relevance to Murex, and providing managers with up-to-date reports of the skills of their domains.
FADEL implemented a powerful solution for tracking content usage on the internet and validating rights information and expiration dates for the content. Content Eye features an internet image indexer which scans specified internet or social media sites to find all images published on a site, and a matching algorithm that identifies any matches to the image in the client asset database. In this document, we present the current capabilities of the systems and its current limitations. We also outline the intended goals of the project by listing the incremental capabilities that we want to add in image processing/matching fields.
Come up with a solution that increments the image match library capability to support some of the image manipulations that are relevant to the use case that Content Eye supports.
The ultimate solution/library would offer APIs that takes the source image as a parameter and return matching images and their score by searching the underlying image repository. If open source libraries are identified to be helpful to the overall implementation, FADEL allows using these libraries after doing assessment on their license model and maturity. A containerized solution using Docker is acceptable as a final deliverable.
As we periodically scan customer's home network to provide assistance, how can we identify what connected devices exist? What are some valuable insights we can proactively provide to users? Is their network vulnerable to some attacks? What would be useful for the customer to proactively know? Can we detect interpret-able anomalies in the network?
Use of non-WiFi and non-3G comms system in case of emergencies (when no WiFi and no 3G exist) This initiative is to find a means to allow AUB to communicate with its constituents in case of emergencies and when there is no WIFI or 3G. Creating a prototype of a working system is required.
Use of drones to improve process efficiency. Chose one of the following initiative of using drones to improve our business process
From audio recording data, identify noisy environment (car, kids, wind, etc.) in real time and suggest customers to switch from voice to digital/messaging channels. Identify poor signal/connection during a voice call and recommend digital experiences. This is not only a noisy channel detection problem, it is also a noise classification problem. The switch suggestion might be: “Kindly drive safely and connect later at this link … Your place in the line is preserved.”
Original PDF from company with 10 FYP ideas:
The time domain reflectometer (TDR) and vector network analyzer (VNA) are the staple instruments to consider, each one having its pros and cons. These two tools are not indiscriminately interchangeable due to the fact that TDR works in the time domain while the VNA functions in the frequency domain. However, the use of a fast Fourier transform (FFT) allows us to use either the two instruments for time and frequency domain measurements.
Multilane has designed multiple TDR boards units, like ML4025-TDR, and ML4035.
Accurate calibration technique for the characterization of one or two port networks using time domain Reflection/Transmission (TDR/T) is needed for these units to get accurate measurements while comparing to a known well calibrated reference like as 50 GHz 4-ports Keysight’s VNA adopted by Multilane.
This project consist of developing, based on raspberry-pi or Arduino, a software along with complementary sensors and electronic board to measure, read and generate DC and AC Current/voltages, temperature, input/output signals…. And to communicate via HTTP (for logging transmission) and SNMP (for instantaneous alarms reporting) to an existing platform along with remote upgrade of the software. Project’s main tasks:
This project consist of communicating with local devices (at sites) that don’t have history logging features. The software should be developed on Arduino or raspberry-pi to log into the device collect all the needed measurements and store them on a CSV file. Any value drifted from predefined threshold will trigger a respective SNMP trap to the server. Also the software should allow an HTTP communication to extract the CSV file by the server/platform as well as remote software upgrade. Project’s main tasks:
Outdoor cabinets have many advantages in telecom industry. However, they are challenging when it comes to cooling. Therefore, this project aims to develop an electronic controller that drives up to 6 PWM fans (pre-selected) based on the internal temperature. The fans and sensors will be installed in different compartments so sensors should be assigned to specific fans to allow correct operation (software assignment is preferred). Again, communication via SNMP (for instantaneous alarms) and HTTP (for historical logging) are needed along with remote software upgrade.
This project aims to assess the technical and commercial feasibility of building a lithium ion battery block of 48V. the battery block should be able to work autonomously without any communication with external charger or parallel battery. However, the battery should be able to communication via HTTP, SNMP along with remote software upgrade. Project’s main tasks:
VRLA Batteries has always been a valuable commodity and thus the vandalism rate is always high. A lot of approaches/solutions have been developed to overcome this issue, but they all had a limited impact on losses due to battery vandalism. This project consist on developing a battery structure approach that will inhibit the reason of stealing the batteries (Lead value, using the batteries in automotive and UPS application for homes…). The students are welcome to think outside the box for innovative approaches on the electrical level, battery shape, packaging or any other means that the students prove valid.
Professors Riad Chedid and Naseem Daher agreed on the following two FYP concepts with Liban cables. Interested professors should discuss their proposals further with the company.
Piloting “water capture from the air for use on campus” – there are many options out there, and the trick would be who to integrate that into our infrastructure. The solution should be green and use renewed energy. Part of the project should also be to study the water usage at AUB and find ways to improve efficiency
Analyze IT Helpdesk service tickets we got over the past couple of years and identify opportunities to automate. Create a Chatbot for IT helpdesk taking into consideration the analysis made and making it actionable to take over some of the manual work we do (Ex reset password, enable AUBNET account, etc..)
It will also be great of we can create a learning mechanism of the chatbot so that it become smarter as we move along.
Dr Najat Saliba has been working on an initiative to study the air particles collected through sensors in Ras Beirut. Use machine learning (Unsupervised) to classify and cluster the data in order to create new discoveries. In addition create data visualization to the data collected to help analyze the data and create value.
IT has created multiple on premise custom solutions that vary from financial to academic administrative solutions. Many of these solutions could be commercialized and made available to other Universities. This project is about creating the framework needed to help transform these solution to a cloud base multi tenant solution. This should include the technical, functional and business model to be used to deploy such solutions.
To help make the university certificate available to public through a secured and reliable platform, we are considering integrating these transcripts with linked in using Blockchain. This will allow students to post their transcripts on their linked in account when they graduate so job seekers will be able to easily access in a secure and controlled manner.
Improve digital compliance by creating a system that can go through our published digital content (AUB Website, Social Media, etc..) and identify names and vet them through the black list. The system could use AI and Data mining techniques to help vet our content and highlight any compliance issue.
Use of AI to help the section capacity planning at AUB. This requires looking at the capacity and sections scheduled over the past couple of years and use machine learning techniques to help predict the need for the coming semester per course.
The idea here is to create a means for students to consult with virtually to get answers to their questions related to a course content. This solution should integrate with Moodle which is our LMS tool. (Rana can provide additional requirements)
FADEL, innovator of rights and royalty management software, has worked with some of the biggest names in media, life sciences, fast-moving consumer goods, publishing, high-tech, and advertising. By automating talent and content rights management across videos, photos, ads, music, products, and brands, and streamlining the processing of licensing royalties, FADEL’s cloud-based solutions have empowered businesses to significantly maximize revenues and increase process efficiencies. Founded in 2003, FADEL is headquartered in New York City and also operates in Los Angeles, London, Paris, and Lebanon. For more information, please visit https://fadel.com/company/
About Rights Cloud Content Eye
https://fadel.com/fadel-announces-general-availability-of-content-eye/
Videos https://fadel.com/resources-demos-webcasts-podcasts/
Rights Cloud for Advertising Content Rights Management
https://www.youtube.com/watch?v=_Wa9TjqRjjo&feature=youtu.be
Rights Cloud Content Eye for Content Tracking and Compliance
Our tech help: service is the easiest way to get more from your tech.
We provide easy, instant access to a community of tech experts, to solve any technology issue you could ever have with almost any device you’ll ever own.
Protect your phone from loss, theft and more with a smartphone insurance plan.
Extended protection plans for TVs, laptops, tablets, home appliances, and more
For detailed Info visit https://www.asurion.com/
Outsource your customer service
High-quality customer service for high-growth online retailers. Our US-based, on-demand 24/7 customer service outsourcing solution can be up and running for your store within hours. 1
Simplr's platform uses your knowledge base, macros, and historical customer inquiries to understand your business and brand.
Our highly-trained and vetted customer support team is on standby to receive your inbound emails, live chats, and/or SMS messages.
We use machine learning and AI to assist Simplr support specialists and aim to resolve 50%+ of customer inquires.
Your customers get the answers they need quickly and accurately. You get better first response time and higher customer satisfaction. Even during high-volume times
For detailed Info visit https://www.simplr.ai/