Programs 2018

Schedule 2018

Workshop is at the convention Center Room 520

Time Event Speaker Institution
09:00-09:10 Opening Remarks BAI
09:10-09:45 Keynote 1 Yann Dauphin Facebook
09:45-10:00 Oral 1 Sicelukwanda Zwane University of the Witwatersrand
10:00-10:15 Oral 2 Alvin Grissom II Ursinus College
10:15-10:30 Oral 3 Obioma Pelka University of Duisburg-Essen Germany
10:30-11:00 Coffee Break + poster
11:00-11:35 Keynote 2 Ayanna Howard Georgia Institute of Technology
11:35-11:50 Oral 4 Randi Williams MIT Media Lab
11:50-12:05 Oral 5 Justice Amoh Dartmouth College
12:05-12:20 Oral 6 Kehinde Owoeye University College London
12:20-02:15 Lunch + Poster
02:15-02:30 Oral 7 Lucio Dery Facebook Inc. / Stanford University
02:30-02:45 Oral 8 Inioluwa Deborah Raji University of Toronto
02:45-03:00 Oral 9 Tewodros Gebreselassie Addis Ababa University
03:00-03:15 Oral 10 Raesetje Sefala Wits University
03:15-03:45 Coffee Break + poster
03:45-04:20 Keynote 4 Brittny-Jade Saunders NYC Commission on Human Rights
04:20-04:55 Keynote 5 Terrence Wilkerson Entrepreneur
04:55-05:25 Panel Discussion on AI Ethics Ezinne Nwankwo (moderator), Stephanie Dinkins, Ayanna Howard, Brittny Saunders, and Terrance Wilkerson
05:25-05:30 Closing Remarks
Dinner Schedule (HYATT REGENCY MONTRÉAL)
07:00-07:30 Reception
07:30-08:00 Welcome to Dinner
08:00-10:00 Dinner
08:30-08:45 BAI presentations
08:45-09:00 Keynote 1 Stephanie Dinkins Stony Brook University, Data & Society
09:00-09:15 Keynote 2 Karim Beguir InstaDeep
09:15-09:30 Keynote 3 Vukosi Marivate University of Pretoria / CSIR
10:00-02:00 am Dancing/Music

Keynote Speakers


Ayanna Howard

Title: “Ensuring a Better World through Engineering, AI and Yes - ROBOTS”

Bio: Ayanna Howard, Ph.D. is the Linda J. and Mark C. Smith Professor and Chair of the School of Interactive Computing at the Georgia Institute of Technology. Dr. Howard’s career focus is on intelligent technologies that must adapt to and function within a human-centered world. Her work, which encompasses advancements in artificial intelligence (AI), assistive technologies, and robotics, has resulted in over 250 peer-reviewed publications in a number of projects - from healthcare robots in the home to AI-powered STEM apps.  To date, her unique accomplishments have been highlighted through a number of awards and articles, including highlights in USA Today, Upscale, and TIME Magazine, as well as being recognized as one of the 23 most powerful women engineers in the world by Business Insider. In 2013, she also founded Zyrobotics, which develops STEM educational products to engage children of all abilities.


Brittny-Jade Saunders

Title: “Local Government & the Challenge of Algorithmic Accountability “

Bio: Brittny Saunders is Deputy Commissioner for Strategic Initiatives at the NYC Commission on Human Rights (“Commission”). At the Commission, Brittny manages key inter-agency partnerships and special projects related to data-driven discrimination and racial justice among others. Before joining the Commission, Brittny worked for the Office of the Mayor, most recently as Acting Counsel to the Mayor. Prior to that, as Deputy Counsel, Brittny played a central role in the Office’s broadband equity efforts, working to ensure affordable access to high-speed internet for residents of the five boroughs. Before joining local government, Brittny worked for the Center for Popular Democracy (“CPD”), where she was Supervising Attorney for Immigrant Rights and Racial Justice, and as Senior Advocate at the Center for Social Inclusion (“CSI” now “RaceForward”). Brittny was a 2010 Fellow in Media, Information & Communications Policy with the Rockwood Leadership Institute and a 2018 Wasserstein Fellow at Harvard Law School. She has an A.B. from Harvard College and a J.D. from Harvard Law School.

Abstract: Brittny will provide an overview of the work of the New York City Commission on Rights, the New York City Human Rights Law and the agency’s developing work on data-driven discrimination. In addition, she will speak about her work on other local government efforts at the intersection of human rights and emerging technologies.


Karim Beguir

Title: “Life of an ML Startup”

Bio: Karim helps companies get a grip on the latest AI breakthroughs and deploy them. A graduate of France’s Ecole Polytechnique and former Program Fellow at the Courant Institute in New York, Karim has a passion for teaching and using applied mathematics.

This led him to launch InstaDeep, a fast-growing African AI startup that focuses on decision making for the Enterprise. Nominated at the MWC17 in the Top 20 most intriguing startups by PCMAG, InstaDeep now has offices in Tunis, London, Paris and Nairobi.

Karim is also the founder of the TensorFlow Tunis Meetup and a Google Developer Expert in ML. He regularly organizes educational events and workshops to share his experience with the community, including mentoring in ML at Google Launchpad Accelerator Africa. Karim is on a mission to democratise AI and make it accessible to a wide audience.

Stephanie Dinkins

Title:

Bio:


Terrence Wilkerson

Bio: Terrence is a family man and the proud father of four daughters. Born and raised in the Bronx, Terrence first encountered the criminal legal system at a young age. He was twice wrongfully accused of crimes he did not commit: once at 19 and again at 40. He is eager to share lessons from those experiences.


Vukosi Marivate

Title: “Building the bridge”

Bio: Vukosi holds a PhD in Computer Science (Rutgers University) and MSc & BSc in Electrical Engineering (Wits University). He has recently started at the University of Pretoria as the ABSA Chair of Data Science. Vukosi works on developing Machine Learning/Artificial Intelligence methods to extract insights from data. A large part of his work over the last few years has been in the intersection of Machine Learning and Natural Language Processing(due to the abundance of text data and need to extract insights). As part of his vision for the ABSA Data Science chair, Vukosi is interested in Data Science for Social Impact, using local challenges as a springboard for research. In this area Vukosi has worked on projects in science, energy, public safety and utilities. Vukosi is an organizer of the Deep Learning Indaba, the largest Machine Learning/Artificial Intelligence workshop on the African continent, aiming to strengthen African Machine Learning. He is passionate about developing young talent, supervising MSc and PhD students and mentoring budding Data Scientists.


Yann Dauphin

Title:

Bio: His research focuses on understanding and developing deep learning algorithms. These algorithms are already helping make our lives better, but they could also help us understand ourselves. He is interested in creating deep learning algorithms that can learn with little supervision and to understand the principles of learning. He completed his Ph.D. at U. of Montreal with Yoshua Bengio on the subject of scaling deep learning algorithms. His collaborators and him have won two international AI competitions: the Unsupervised Transfer Learning Challenge in 2011, and the EmotiW challenge in 2014.

Oral Research Presenters


Alvin Grissom II

Title: “Pathologies of Neural Models Make Interpretations Difficult”

Abstract: One way to interpret neural model predictions is to highlight the most important input features—for example, a heatmap visualization over the words in an input sentence. In existing interpretation methods for NLP, a word’s importance is determined by either input perturbation—measuring the decrease in model confidence when that word is removed—or by the gradient with respect to that word. To understand the limitations of these methods, we use input reduction, which iteratively removes the least important word from the input. This exposes pathological behaviors of neural models: the remaining words appear nonsensical to humans and are not the ones determined as important by interpretation methods. As we confirm with human experiments, the reduced examples lack information to support the prediction of any label, but models still make the same predictions with high confidence. To explain these counterintuitive results, we draw connections to adversarial examples and confidence calibration: pathological behaviors reveal difficulties in interpreting neural models trained with maximum likelihood. To mitigate their deficiencies, we fine-tune the models by encouraging high entropy outputs on reduced examples. Fine-tuned models become more interpretable under input reduction without accuracy loss on regular examples.


Inioluwa Deborah Raji

Title: “In the Shadow of Gender Shades”

Abstract: While there have been mounting calls for algorithmic transparency as more artificial intelligence services become mainstream, and audit approaches for online platforms have been proposed, audit strategies for the effective design and disclosure of external evaluations of commercial pretrained machine learning models distributed as Application Program Interfaces (APIs) remains underdeveloped. This paper thus extends scholarship on the development and impact of black-box algorithmic auditing by exploring a real-world commercial facial analysis intersectional audit and the corporate reactions to the audit release. This paper 1) outlines the audit design and the public and private audit disclosure procedure used in the Gender Shades study, 2) presents new performance metrics from originally targeted companies T-1, T-2, T-3 on the Pilot Parliaments Benchmark as of August 2018, 3) provides performance results on PPB by non-target companies NT-A and NT-B and 4) Explores differences in company responses as shared through corporate communication.


Justice Amoh

Title: “An Optimized Recurrent Unit for Ultra-Low Power Acoustic Event Detection”

Abstract: Currently, there is a demand for neural network models that can run on sensors, wearables and IoT devices. However, resource constraints of such edge devices make it challenging to realize on-device neural network inferencing. For ultra-low power wearable applications, there are no practical solutions for deploying neural networks. To meet this need, our work introduces a new recurrent unit architecture that is specifically adapted for on-device low-power acoustic event detection (AED). The proposed embedded Gated Recurrent Unit (eGRU) is based on the GRU architecture but features optimizations that make it implementable on ultra-low power micro-controllers such as the ARM Cortex M0+. With our proposed modifications, eGRU is demonstrated to be effective, especially for short duration AED and keyword spotting tasks. A single eGRU cell is 60x faster and 12x smaller than a GRU cell. Despite its optimizations, eGRU compares well with conventional GRU across AED tasks of different complexities.


Kehinde Owoeye

Title: “Identifying sheep with abnormal movement trajectory in a flock”

Abstract: Learning to identify anomalies in the behavior of individuals is becoming increasingly important for a variety of reasons for example in studying the progression of several diseases. Due to the need to assess the efficacy of therapeutic interventions, animals with longer life span are becoming increasingly important for assessing the efficacy of therapeutic interventions. In this presentation, I will describe computational methods that allow for the automatic discrimination of sheep with a genetic mutation that causes Batten disease from an age-matched control group, using GPS movement traces as input. Batten disease is a neurodegenerative disease with symptoms that are likely to affect the way that those with it move and socialize, including loss of vision and dementia. The distance covered in each ten minute period and, more specifically, outliers in each period, are used as the basis for identification. Our results show that, despite the variability in the sample, the bulk of the outliers during the period of observation came from the sheep with Batten disease. Our results, though preliminary, point towards the potential of using relatively simple movement metrics in identifying the onset of a phenotype in symptomatically similar conditions.


Lucio Dery

Title: “Audio to Body Dynamics”

Abstract: We present a method that gets as input an audio of violin or piano playing, and outputs a video of skeleton predictions which are further used to animate an avatar. The key idea is to create an animation of an avatar that moves their hands similarly to how a pianist or violinist would do, just from audio. Aiming for a fully detailed correct arms and fingers motion is a goal, however, it’s not clear if body movement can be predicted from music at all. In this paper, we present the first result that shows that natural body dynamics can be predicted at all. We built an LSTM network that is trained on violin and piano recital videos uploaded to the Internet. The predicted points are applied onto a rigged avatar to create the animation.


Obioma Pelka

Title: “Radiology Objects in COntext (ROCO): A Multimodal Medical Image Dataset”

Abstract: This work introduces a new multimodal image dataset, with the aim of detecting the interplay between visual elements and semantic relations present in radiology images. The objective is accomplished by retrieving all image-caption pairs from the open-access biomedical literature database PubMedCentral, as these captions describe the visual content in their semantic context. The target domain being radiology, all compound, multi-pane, and non-radiology images were eliminated using an automatic binary classifier fine-tuned with a deep convolutional neural network system and trained with datasets distributed at the medical tasks of ImageCLEF 2013, 2015 and 2016. The Radiology Objects in COntext (ROCO) dataset contains over 81k radiology images with several medical imaging modalities including Computer Tomography (CT), Ultrasound, X-Ray, Fluoroscopy, Positron Emission Tomography (PET), Mammography, Magnetic Resonance Imaging (MRI), Angiography and PET-CT. For all images in ROCO, the corresponding caption, keywords, Unified Medical Language Systems (UMLS) Concept Unique Identifiers (CUIs) and Semantic Type will be distributed. An additional out-of-class set with 6k images ranging from synthetic radiology figures to digital arts is provided, to improve prediction and classification performance of out-of-class samples. Adopting ROCO, systems for caption and keywords generation can be modeled, which enables multimodal representation for image datasets lacking text representation. Furthermore, systems with the goal of image structuring and semantic information tagging can be created using ROCO, which is beneficial and of assistance for image and information retrieval purposes.


Raesetje Sefala

Title: “Using satellite images and computer vision to study the evolution and effects of spatial apartheid in South Africa”

Abstract: One of the main problems in South Africa is removing many of the legacies of Apartheid - a former policy of economic and political discrimination, and segregation against non-European groups in South Africa. For example, moving around residential areas shows the legacy of spatial apartheid on a smaller scale- completely segregated neighborhoods of townships next to gated wealthy neighborhoods that have largely remained unaffected by the ending of apartheid. Our research proposes to use computer vision to analyze millions of such satellite images of South Africa from 2006 to 2016. This work aims to use satellite images, geographically labelled coordinates of South Africa’s built environment and socioeconomic data to understand the relationship between the spatial and socioeconomic makeup of neighborhoods in South Africa, and study how they have evolved over time post Apartheid. We propose a semantic segmentation model to detect and classify clusters of townships and wealthy areas from these high resolution satellite images. In addition to automatically detecting and classifying neighborhoods, we plan to use demographic and socioeconomic data to then analyze the change over time in the demographic makeup, economic status and access to basic resources such as the number of clinics and schools of these detected neighborhoods.


Randi Williams

Title: “PopBots: Leveraging Social Robots to Aid Early Childhood Artificial Intelligence Education”

Abstract: Artificial intelligence (AI) is revolutionizing our lives, impacting the way that even the youngest members of society live, learn, and play. Previous work examining children’s relationships with AI has shown that this population lacks an understanding of how AI devices work. This lack of understanding makes it difficult for children to engage in safe and constructive interactions with their smart playthings. Furthermore, as this technology becomes more pervasive, we must think about how to build a diverse workforce that creates technology to equitably address the needs of many. Given these motivations, we designed an early childhood AI curriculum, PopBots. PopBots is a hands-on toolkit that enables young children to learn about AI by programming and training a social robot. We evaluated the toolkit with 80 preschool children (ages 4-6) and found that the use of a social robot as a learning companion and programmable artifact was effective. Children could correctly answer questions about knowledge-based systems, supervised machine learning, and generative music algorithms. Additionally, we found that using the toolkit helped children better appreciate the cognitive capabilities of robots. We will discuss the toolkit and teaching methods used in hope that this first exploration into early AI education will inspire other educators and researchers.


Sicelukwanda Zwane

Title: “Safer Exploration in Deep Reinforcement Learning using Action Priors”

Abstract: Behavior learning in deep reinforcement learning is inherently unsafe because untrained agents typically have to sample actions from randomly initialized task policies and from random exploration policies. Executing these actions in physical environments can lead agents to harmful states, possibly causing damage and poor initial performance. In this work, we address this problem by using transfer learning to develop a framework for safer reinforcement learning in continuous environments. We show that our exploration policy results in fewer collisions with the environment, better initial performance, and earlier convergence compared to the vanilla epsilon-greedy random exploration policy.


Tewodros Abebe Gebreselassie

Title: “Parallel Corpora for bi-Directional Statistical Machine Translation for Seven Ethiopian Language Pairs”

Abstract: In this paper, we describe the development of parallel corpora for Ethiopian Languages: Amharic, Tigrigna, Afan-Oromo, Wolaytta and Ge’ez. To check the usability of all the corpora we used them to conduct a baseline bi-directional statistical machine translation experiment for seven language pairs. The bi-directional SMT BLEU score shows that all the corpora can be used for further SMT investigations. We have also learnt that the morphological richness of the Ethio-Semitic languages has a negative impact on the performance of the SMT especially when they are target languages. Now we are working towards selecting an optimal alignment for bi-directional statistical machine translation among the Ethiopian languages.

Accepted Posters

Name of presenter Title of Poster
Bitseat Tadesse Aragaw Sentence Level Amharic Text Sentiment Analysis Model: A Combined Approach
Bitseat Tadesse Aragaw Sentence Level Amharic Text Sentiment Analysis Model: A Combined Approach
Wathela Alhassan The FIRST Classifier: compact and extended radio galaxy classification using deep Convolutional Neural Networks
Neema Mduma Evaluation of Imbalanced Data Techniques for Student Dropout Prediction
Neema Mduma Evaluation of Imbalanced Data Techniques for Student Dropout Prediction
Oyebo Abdulhamiid Bankole PLANT LEAVES CLASSIFICATION USING K-NEAREST NEIGHBOUR ALGORITHM
Geletaw Sahle Tegenaw A machine learning approach for identifying proteins involved in conserved parasite- mosquito interactions
Bruno Ssekiwere A Blended Approach of Machine Learning Techniques in Predicting Vegetation Cover
Semakula Abdumajidhu Sootymold effect on cassava yields using convolutional neural networks.
Semakula Abdumajidhu Sootymold effect on cassava yields using convolutional neural networks
Omotola Dawodu Application of Machine Learning to Classification of Diabetes Mellitus
Mulubrhan Hailegebreal A Bidirectional Tigrigna – English Statistical Machine Translation
BABALOLA Moyin Florence Condition-Based Knowledge Representational Structure for Identifying Norms Violation In Logic-Based Normative Systems
Abiodun Modupe Deep Learning for Authorship Attribution of Social Media Forensics
Zelalem Fantahun Abate Unsupervised Part-of-Speech Tagger for Amharic using K-means clustering
Abel Kahsay Deep learning Based Gastrointestinal Disease Recognition for Endoscopic Images
Hafte Abera Design of Tigrinya Language Speech Corpus for Speech Recognition
Emeka Ogbuju Development of a Deep Sentiment Recommender for E-commerce
Hicham Hammouchi Visual Speech Recognition using Hahn Convolutional Neural Networks
Hicham Hammouchi Visual Speech Recognition using Hahn Convolutional Neural Networks
Michael Melese Woldeyohannis English-Ethiopian Languages SMT
Alemayehu Solomon Admasu Deep Haar scattering networks in pattern recognition: a promising approach
Kehinde Aruleba Automatic Recognition of Hand-drawn Finite Automata Images
Kehinde Aruleba Recognition of hand-drawn finite automata images
Blessing Ogbuokiri Examining Social Media Impact on the Politics of Nigeria Using Social Network Analytics
Blessing Ogbuokiri Examining Social Media Impact on the Politics of Nigeria using Social Network Analytics
Gereziher Weldegebriel Adhane Multiple face detection and tracking in a real-time video sequences using centroid tracking and deep learning techniques
Abraham Enyo-one Musa EXPERT SYSTEM FOR EYE DISEASE DIAGNOSIS
MUSA, Abraham Enyo-one EXPERT SYSTEM FOR EYE DISEASE DIAGNOSIS
selam waktola Automatic stagnant zone segmentation using CNN and x-ray tomography of silo discharging process
selam waktola Automatic stagnant zone segmentation using CNN and x-ray tomography of silo discharging process
Jeraldy Deus TRANSFER LEARNING APPLIED TO BANKNOTE RECOGNITION FOR VISUALLY IMPAIRED PEOPLE.
George Rabeshi Obaido Automatic Plagiarism Detection in Student Programs using PlaGraph
George Rabeshi Obaido Automatic Plagiarism Detection in Student Programs using PlaGraph
Sakinat Folorunso EMPIRICAL COMPARISON OF TIME SERIES DATA MINING ALGORITHMS
Lindelweyizizwe Manqele Smarter decision-making using Internet of Things enabled sensor data
PRINCE MAKAWA ABUDU Communicating Recurrent Neural Networks for Resource Constrained Systems
Prince M Abudu Communicating Recurrent Neural Networks for Resource Constrained Systems
Chidubem G Arachie Adversarial Learning for Weak Supervision
Abdullah Khadijha-Kuburat Adebisi A PREDICTIVE MODEL FOR TWEET SENTIMENT ANALYSIS AND CLASSIFICATION
Babirye Claire Mining for Votes: Inferring Voting Trends from Twitter Data
George Musumba Towards an IT-Mediated Food Insecurity Solution for Developing Nations
George Musumba Towards an IT-Mediated Food Insecurity Solution for Developing Nations
Natasha Williams Medical Artificial Intelligence: The Inclusion of Racial and Ethnic Minorities in Clinical Trials Will Improve Data Diversity
Natasha H. Williams, PhD, JD, LLM, MPH Medical Artificial Intelligence: The Inclusion of Racial and Ethnic Minorities in Clinical Trials Will Improve Data Diversity
Zimele Gwebu Portable Pedestrians and Animals Detection Device for Vehicles using Transfer Learning
Omara Patrick MULTI-RISK ANALYSIS OF PROSTATAE CANCER SURVIVAL
Omara Patrick MULTI_RISK ANALYSIS OF PROSTATE CANCER SURVIVAL
Omara Patrick Multi-Risk Analysis of Prostate Cancer Survival
Tameru Hailesilassie Rule Extraction Algorithm for Deep Neural Networks: A Tool Towards Explainable AI
Tameru Hailesilassie Rule Extraction Algorithm for Deep Neural Networks: A Tool Towards Explainable AI
Solomon Nsumba Automated image-based diagnosis of cowpea diseases
Melles, Abey Desta Unsupervised Similarity Based Topic Segmentation System for Amharic
Israel Goytom A Machine learning approach to detect and classify 3D two-photon polymerization microstructures using optical microscopy images
Israel Goytom A Machine learning approach to detect and classify 3D two-photon polymerization microstructures using optical microscopy images
Idowu T. Aruleba Hypertension Prediction System Using Naive Bayes Classifier
Omar Transfer Reinforcement Learning Through Demonstration
Adesina adetola PREDICTING RATE OF ACCEPTANCE AND USE OF WEB PORTAL SERVICE (UNIVERSITY ENVRIONMENT) USING MULTILAYER PERCEPTRON FEEDFORWARD ARTIFICIAL NEURAL NETWORK
Adesina Adetola Sunday Predicting rate of acceptance of web services (university environment) using multilayer perceptron feedforward artificial neural network
Adesina Adetola Sunday Predicting rate of acceptance and use of web portal service (university environment) using multilayer perceptron feedforward artificial neural network
kiante brantley Learning to Teach: Learning Good Teaching Policies with Reinforcement Learning
Kiante Brantley Learning to Teach
Femi Alayesanmi Samson MACHINE LEARNING IN FORENSICS: DEVELOPING AN OBJECT DETECTION MODEL FOR CRIME EVIDENCE ANALYSIS USING YOLO
Alayesanmi Femi Samson Object Detection Model for Crime Evidence Analysis Using Yolo
Irene Nandutu Building a Neural Machine Translation with Attention for Low Resource Language – Luganda
Kyamanywa Kenneth Dynamic Route Optimization for Public Trasportation Using Crowd Sourced User Feed
Kenneth Kyamanywa Dynamic Route Optimisastion
Simphiwe Zitha Classifying Radio Galaxies using Model-Agnostic Meta-Learning
Simphiwe Zitha Classifying Radio Galaxies using Model-Agnostic Meta-Learning
Mutembesa Daniel Crowdsourcing real-time disease and pest information. A case of nation-wide cassava disease surveillance in a developing country.
Bereket Abera YILMA Constructive Social Choice with Setwise Max-margin
Melese Mihret Wondim Sentiment Analysis model for opinionated Awngi text: case of Music reviews
Melese Mihret Wondim Sentiment Analysis model for opinionated Awngi text: Case of Music reviews
Odu Nkiruka Bridget A Fuzzy-based Approach for Modelling Preferences of Users in Multi-criteria Recommender Systems
VICTOR C DIBIA COCO-Africa: A Curation Tool and Dataset of Common Objects in the Context of Africa
Robert Nsinga Predictive Analysis of Cohesiveness in Multivariate Sequences Using LSTM Recurrent Neural Networks
Elhadji Amadou Oury DIALLO Learning Group Formation for Coordinated Behavior in Adversarial Multi-Agent with Double DQN
Kibrewossen Yitbarek Mekasha AMHARIC SENTENCE GENERATION FROM INTERLINGUA REPRESENTATION
Kibrewossen Yitbarek Mekasha Amharic sentence generation from Interlingua representation
Devotha Nyambo Modeling of Annual Milk Yield of Dairy Cows: Impact of Exclusive Grazing and Infectious Disease to Individual Cows
Muthoni Wanyoike Promoting diversity and inclusion in the field of Artificial Intelligence in Kenya
Isaac Mukonyezi Prediction of Spectrum Holes in Cognitive Radio Ad-Hoc Networks
Taiwo Abass Ishola Modelling and Forecast Evaluation Performance with Extended Neural Network using Climatic Time Series Data
Thon Kuir Biar Ayual AGRICULTURAL ADVISORY ON SEASONAL VARIATION IN SOUTH SUDAN
Thon Kuir Biar Ayual Computational Weather prediction in South Sudan
Randi Williams PopBots: Leveraging Social Robots to Aid Early Childhood Artificial Intelligence Education
Thon Kuir Biar Ayual Computational weather prediction in South Sudan
Masresha Beniam A Classical Method for Detecting Overlapping Faces in Images
Hafte Abera Speech Recognition for Tigrinya language Using Deep Neural Network Approach
Odu Nkiruka fuzzy-based approach in modeling preferences of users in multi-criteria recommender system
Odu Nkiruka Bridget A Fuzzy-based Approach for Modelling Preferences of Users in Multi-criteria Recommender Systems
Bayeleygne Meseret Dastaw Speculative Scientific Inference via Synergetic Combination of Probabilistic Logic and Evolutionary Pattern Recognition
Ahmed Mohammed Y-NET: A DEEP CONVOLUTIONAL NEURAL NETWORK TO POLYP DETECTION
Zekarias Tilahun Kefato Network-Agnostic Cascade Prediction in Online Social Networks
Darlington Ahiale Akogo CellLineNet: End-to-End Learning and Transfer Learning For Multiclass Epithelial Breast Cell Line Classification Via A Convolutional Neural Network
Abdullah Mohamed, Abubakr Hassan Implementation of A Neural Natural Language Understanding Component for Arabic Dialogue Systems
Abdallah Mohammed/ Abubakr Hassan Implementation of a neural natural language understanding for Arabic dialogue system
juvenalis musungu Optimizing Agricultural Yields using AI for weed control
Irené Tematelewo Blood Glucose-Insulin Metabolism Modeling in Type 1 Diabetics using System Identification
Mohammed Khalil Speaker identification for aeronautical communications systems based on SMFCC and i-vector
Alvin Grissom II Pathologies of Neural Models Make Interpretations Difficult
Basiliyos BETRU Embedding user geo-spatial preference for personalized business opportunity recommendation.
Basiliyos BETRU Embedding Geo-spatial preference for optimal recommendation
Simon Mekit GHOST - A Dialogue and Behavior Scripting framework for Robots and other Intelligent Agents
Simon Mekit A ChatBot Framework for Robots and other Intelligent Agents
Yemisrach G Nigatie Investigating Unsupervised Approach for Amharic Part of Speech Tagging
Randi Williams PopBots: Leveraging Social Robots to Aid Preschool Children’s Artificial Intelligence Education
Felipe Paula Detecting neuropsychiatric conditions with semantic verbal fluency
Felipe Paula Detecting neuropsychiatric conditions with semantic verbal fluency
Olaniyan Oluwasegun Emmanuel DEVELOPMENT OF MULTI-TARGET REGRESSION MODELS TO PREDICT THE PHYSICAL AND CHEMICAL PROPERTIES OF SOIL
Tejumade Afonja CHOWNET: An Image Local Food Dataset
Tejumade Afonja CHOWNET: An Image Dataset For Local Food
Daniel Melesse A Data-Driven Approach to Automatic Gaze Tracking
Francisca Oladipo The Machine Learning of Women: Dataset and Initial Results
Mellitus Okwudili Ezeme Hierarchical Attention-Based Anomaly Detection Model for Embedded Operating Systems
Martha Yifiru Tachbelie Development of Pronunciation Lexicons for Amharic Automatic Speech Recognition (ASR)
Solomon Teferra Abate Development of Pronunciation Lexicons for Amharic Automatic Speech Recognition (ASR)
Martha Yifiru Tachbelie Development of Pronunciation Lexicons for Amharic Automatic Speech Recognition (ASR)
Marcellin Atemkeng Dimensional Reduction Techniques for Radio Interferometric Big Data Compression
ONALETHATA INNOCENT MASWABI Real Time PID Feedback Control Online Tuning Algorithm
Onalethata Innocent Maswabi Real time PID feedback Control Online Tuning Algorithm
Meareg Hailemariam Real-time Mirroring: Human Facial Expressions to a 3D Avatar
Mohammed Khalil Impact of digital watermarking on MR-Brain pathological detection system
Joel Eyamu Predicting Multi-drug resistant tuberculosis using machine learning
Ahmed Elsiddieg Abdulaziz Abdalla A comparative review of incremental learning of sensorimotor models for developmental robots
Obioma Pelka Radiology Objects in COntext (ROCO): A Multimodal Medical Image Dataset
Frederick Apina Leveraging Machine Intelligence for Diagnosing UTI
Bizuayehu Improving SMT Perfomance by Extracting Parallel Sentences from Comparable Corpora for low resourced languages
Yosi Shibberu AI for Africa: Opportunities and Challenges
Yosi Shibberu AI for Africa: Opportunities and Challenges
Habiba Sultan Rega DESIGN AND IMPLEMENTATION OF AN INTEGRATED REAL-TIME BUS AND TRAIN TRACKING FRAMEWORK
Habiba Sultan Rega and Obsa Taera Deressa DESIGN AND IMPLEMENTATION OF AN INTEGRATED REAL-TIME BUS AND TRAIN TRACKING FRAMEWORK
Tlamelo Makati A New Metric for Scoring Video Action Segmentation Methods in a Supervised Setting
Emmanuel Masabo A Self Adaptive Model for Detecting Polymorphic Malware
Mpho Mokoatle Collision Course: Challenges with Road Traffic Accident Data in South Africa
owoeye kehinde Identifying sheep with abnormal movement trajectory in a flock
Kehinde Owoeye Identifying sheep with abnormal movement trajectory in a flock
Vongani Maluleke Aerial Image Poverty Estimation
Vongani Maluleke Aerial Image Poverty Estimation
Bayanda Chakuma Visualizing the Optimization process for Multi-Objective Optimization Problems
Bayanda Chakuma VISUALIZING THE SEARCH PROCESS FOR MULTI-OBJECTIVE OPTIMIZATION PROBLEMS
Bayanda Chakuma Visualizing the Search Process for Multi objective Optimization Problems
Adam Kyomuhendo LEGALITY AND THE ETHICS OF USE OF UNMANNED COMBAT AERIAL VEHICLES (UCAVs) IN THE CONTEXT OF INTERNATIONAL HUMANITARIAN LAW
Kassahun Tamir Handwritten Amharic Characters Recognition Using CNN
Daniel Nkemelu Deep Convolutional Neural Network for Plant Seedlings Classification
Nofundiso Path planning on a constrained manifold
El Mehdi El Allaoui Empirical Evaluation of Word Representations on Arabic Sentiment Analysis
El Allaoui El Mehdi Empirical Evaluation of Word Representations on Arabic Sentiment Analysis
George Boateng Multimodal Affect Detection among Couples for Diabetes Management
Roy Henha Eyono Learning To Backpropagate
Ermias Abebe Tegegn Machine learning Model for Predicting the Status of HIV Patients during Drug Regimen Change
Ermias Abebe Tegegn Machine learning Model for Predicting the Status of HIV Patients during Drug Regimen Change
Raesetje Sefala Using satellite images and computer vision to study the evolution and effects of spatial apartheid in South Africa
Raesetje Sefala Using satellite images and computer vision to study the evolution and effects of spatial apartheid in South Africa
Mamuku Mokuwe Saliency Overlay Generation Using Bayesian Optimisation
Mamuku Windy Mokuwe Saliency overlay generation using Bayesian optimisation
Andrew Zaldivar Model Cards for Model Reporting
Ahmed Babajide Olanrewaju Analysis of the Adoption of Social Media Tools by Government Agencies in Nigeria using Machine Learning Approach
Lucio Dery Audio To Body Dynamics
Terrell Nowlin Predicting Prostate Cancer Reucurrence Using Computer Vision
Melissa Woghiren Revisiting the Prediction of cis-Regulatory Genomic Elements Using Machine Learning Tools
Dr. Modinah A. O. Abdul Raheem Augmented Reality as a Classroom Engagement Tool in Repositioning Geography Learning in Osun State Nigeria
Ajani, Adedeji Hammed Augmented Reality ad a Classroom Engagement Tool in Repositioning Geography Learning in Osun State Nigeria
Jahkel Robin Neural Sentence Reordering for Simultaneous Machine Translation
Ezinne Nwankwo Perspectives on the Use of Algorithms in the Public Sector
yenatfanta shifferaw bayleyegn Early detection of kidney abnormality using neural network
Taha Merghani Application of The Hilbert-Schmidt Independence Criterion to Lexical Geographical Variation in Lyon, France
Thembani Phaweni Extracting structured information from organisational diagrams
Thembani Phaweni Extracting structured information from organograms and network diagrams
Charles C Onu Undersampling and Bagging of Decision Trees in Analysis of Cardiorespiratory Variability for Extubation Readiness in Extremely Preterm Infants
Wondimagegnhue Tsegaye Tufa Morphological Segmentation Using Encoder-Decoder for Morphologically Rich Languages (MRL)
Daniel Ajisafe Early retinal tissue damage detection using Machine Learning
Seifedin Shifaw Amharic Text Normalization for Higher Level NLP Applications Using Machine Learning Approach
Solomon Teferra Abate Amharic Text Normalization for Higher Level NLP Applications Using Machine Learning Approach
Seifedin Shifaw Mohamed Amharic Text Normalization for Higher Level NLP Applications Using Machine Learning Approach
Tlou Boloka Towards knowledge Tranfer for Model-Based Deep Reinforcement Learning
Shelby Heinecke Crowdsourced PAC Learning under Classification Noise
Fisseha Gidey GEBREMEDHIN Data Visualization for Exploring Comparative Advantages in Multidimensional Economic Data
Ismaël Koné Mondrian Forests with Label Guided Splits
Joshua Patterson RAPIDS: GPU Accelerated Data Science
Reem Elmahdi Analysis, Prediction and Comparison Algorithms For Water Quality Variables
Reem Elmahdi Transfer Learning in Water Quality Variable’s Prediction
Latoya Peterson AI in the Trap
Olasunkanmi, Roseline Olawumi Deep Neural Network Based Approach to Skin Cancer Classification
Olawumi Roseline Olasunkanmi Deep Learning Based Approach to Skin Cancer Detection
Henry Burton Machine-learning (ML) based earthquake damage detection of residential buildings
meryem hagui, ABDELATI EL ASRI Convolutional Neural Networks for Breast Cancer through Invasive Ductal Carcinoma
Taiwo Kolajo Capturing Rich Semantics Implicit in Social Media Streams for Improved Analytics Result
Nalwooga Samiiha Machine Learning Approach for Monitoring of Viral Cassava Disease
Eric Corbett Interactive Machine Learning Heuristics
Ndivhuwo Makondo Towards improving sensorimotor model learning for developmental robots with multi-robot knowledge transfer
Tshepiso Mokoena, Koena Monyai Explaining anomalies via Sequential Feature Explanations and Visualisations
Claudia V. Roberts Quantifying the Extent to Which Popular Pre-trained Convolutional Neural Networks Implicitly Learn High-Level Protected Attributes
Cody Coleman Computationally efficient subset selection for deep learning training
Cody Coleman Efficient Data Selection For Training Deep Networks
Mohamed Hassan Kane Learning to learn how to learn
Salahadin Seid Musa Towards Real-time Multimodal Emotion Recognition for Tele-health using IoT
Girmaw Abebe Tadesse Cross-domain knowledge transfer for wearable sensors
Ally Salim Jr Synthetic Patient Generation: Deep Learning to Generate New Patient Records
Omolayo Olasehinde Stock Price Prediction System using Long Short-Term Memory
Sicelukwanda Zwane Safe Exploration in Deep Reinforcement Learning with Action Priors
Samee Ibraheem Speech Recognition Diversity in Clinical Settings
Linda Khumalo Modelling long-range contextual information in a Recurrent Neural Network Language Model
Abstractive text summarization with attention Abstractive text summarization with attention
Ditebogo Masha Proprioceptive Terrain Classification for Tracked Mobile Robots using SVM
Waleed Khamies & Montaser Mohammedalamen Transfer Learning For Prosthetics Using Imitation Learning
Thabo Malete EEG-based Control of a 3D Game Using a 14-channel BCI
Justice Amoh e-GRU: An Optimized Recurrent Unit for Ultra Low Power Acoustic Event Detection
Justice Amoh An Optimized Recurrent Unit for Ultra-Low-Power Acoustic Event Detection
Inioluwa Deborah Raji In the Shadow of Gender Shades: Case-based Exploration of Corporate Reactions to a Third-Party Black Box Algorithmic Audit
Victoria Okuneye Classification of Psychosis Diagnoses using Resting State Functional Connectivity from Multi-Site Bipolar-Schizophrenia Intermediate Phenotype Study
Christine Allen-Blanchette Design and Use of Equivariant Filters in CNNs
Zimkhitha Sijovu Probabilistic state estimation and calibration for a robot manipulator end-effector.
Amr Khalifa & Eltayeb Ahmed Generating Optimized Traffic Light Controllers using Reinforcement Learning
AnnMargaret Tutu DeepBlock: a decentralized approach to hardware acceleration for deep learning.
Tesfamariam M Abuhay Data-Driven Simulation of Patient Flow through Multiple Departments to Estimate Load of Departments
Hiba Chougad Multi-label Transfer Learning for the Early Diagnosis of Breast Cancer
Ms Martha Shaka IMPROVING PROPERTY TAX COMPLIANCE: MACHINE LEARNING APPROACH
FREDRICK MANANG IMPROVING PROPERTY TAX COMPLIANCE: A MACHINE LEARNING APPROACH
Tuga Abdelkarim Ahmed Detecting Depression and Suicidal Thoughts on Social Media
Flora Tasse 3D Scene Estimation From Images
Flora Ponjou Tasse 3D Scene Estimation From Images
Onyeka Emebo A Conceptual Framework for Team Selection using Semantic Case-based Reasoning
Hope Mogale Training and Optimizing Music Recommendation Algorithms Using Self-Similarity Matrices
Oyebo Abdulhamiid Bankole Plant leaves Classification Using K- Nearest Neighbour
Eric Mibuari Subset Scanning for Anomaly Detection on Customs Data
Babalola Moyin Florence Condition-Based Knowledge Representational Structure for Identifying Norms Violation In Logic-Based Normative Systems
Milen Girma Kebede Regulatory Compliance in Healthcare: Managing Consent
Surafel Melaku Lakew Improving Extremely Low-Resource and Zero-Shot Neural Machine Translation
Koena Monyai Explaining anomalies via Sequential Feature Explanations and Visualisations
Abiodun Modupe Deep Learning framework for Authorship Attribution of Social Media Texts
Michael Melese Woldeyohannis English-Ethiopian Languages Statistical Machine Translation
Victor Dibia COCO-Africa: A Curation Tool and Dataset of Common Objects in the Context of Africa
Sarah Brown Critical Data Exploration by Detecting Simpson’s Paradox
Matthew Tesfaldet Convolutional Photomosaic Generation via Multi-Scale Perceptual Losses