Posters

Poster Session Details

Posters are an interactive and visual way of explaining your work, getting feedback on ideas, and for meeting new people.

  • Over the two days, there will be 154 posters presented in total.
  • The poster session is an informal event where attendees can walk between all posters and exhibitions. It is useful to think about how you can explain the content of your poster very briefly, so that even those who aren’t working in your topic will be able to understand your motivations and learn about a different area.

Poster allocations

Every poster will be presented on either Monday or Tuesday. Find your allocation below.

Session Date Poster Info
Monday September 11th Poster Presentations
First Name Surname Poster Name
AbbottNature inspired swarm robotics algorithms for prioritized foraging
AdebanjoFeature extraction of hyperspectral images
AjoodhaComputationally tracking direct influence in observational data
AtemkengRadio interferometric point spread function morphological distortion: where and why?
BessingerAn integrated coastal map of South Africa
Biyela-
BriersRiver flow routing using machine learning
ChiundidzaBig data deep learning as an alternative for measurement of socioeconomic indicators in developing countries
CollinsActive learning and system identification in accelerated destructive degradation
ConwayAugmenting word embeddings using knowledge base data extracted from Wikipedia
CullinanAgent oriented deep learning
CurrinComputational neuroscience: the brain on AI
De CarvalhoSales forecasting with multivariate linear regression
de LangePlant disease recognition
du PlooyApplying named entity recognition on the dictionary of Southern African place names
DüsterwaldA computational model reveals the roles of impermeant anions and transporters in neuronal Cl- homeostasis
EarleHierarchy through composition with LMDPs
FaustineHybrid HMM-deep learning models for low-sampling energy disaggregation problem
FickGlobal, curve-sensitive features for offline signature verification
GerrandDeep learning for paediatric chest X-ray screening
GilaniGenerative adversarial networks for natural dialogue generation
GilbertNLG for isiZulu sentences using n-grams
GobonamangIntelligent DNA sub sequence repeats identification algorithm
Hamed-
Henha EyonoAuthentication using keystroke dynamics
HosenieSource classification in deep radio surveys using machine learning techniques
JamesLearning portable symbols for high-level planning
JeewaCryptographically secure pseudo-random number generators using artificial neural networks
KhanNeural network training optimization through reinforcement learning
KhumaloLearning context for a deep recurrent neural network language model
KhumaloFacial recognition
LalaUnraveling the contribution of image captioning and neural machine translation for multimodal machine translation
LebeseStock market prediction using artificial neural networks and support vector machines
LiuIncreasing the trading prediction by mining aggregated human texting messages
MabasoSpot detection in microscopy images using convolutional neural network
Malete-
MalobolaIntelligent process automation
MaraisCNNs for multi-label classification
Marom-
MasakunaTackling inconsistency In classifier fusion
MaseleseleHaving vs clustering algorithms for non stationary data streams
McCoyMachine learning applications in minerals processing
MeyerApplying word embedding techniques to medical data
MmopelwaBig data application on investment portfolio management and returns prediction
MpondoComparison of trend for segments of time series data
MsomiData driven approach to business
MthwecuStochastic games
MüllerCompound landmarks for stereo-vision SLAM
MvelaseSecurity surveillance deep learning solutions
NekhumbeBreast cancer lesion detection and segmentation using machine learning techniques
NekotoAn African digital colonial museum: using deep learning, evolutionary psychology and big data tools towards a social transformation/cultural appreciation
NemasisiEvent detection on social media streams
NewmanVideo classification using memory augmented networks
NgejaneMitigating cyber-crime and child grooming on social media using deep learning
NimoAnalysis of galaxy kinematics, dynamics and evolution
NogwanyaFeasibility of nuclear plasma interaction studies with the activation techniques
OluyideUnusual event detection in surveillance videos using deep learning
PerlowTowards a recycling agent using novel material recognition
PetersOptimal image facetting for direction dependent effects
PiseFacial image analysis for e-learning
RajohnsonThe role of dust in star formation
RaseonyanaProducing an optimal examination timetable for the University of Botswana
RensIncorporating learning into an agent's stochastic knowledge management framework
RoySpeech emotion recognition
RozanovaDeep learning in NLP
SeedatQuadcopter control
Sejeso-
SetlhakeDeep learning in multi-spectral image classification
ShabalalaViolence detection and characterization in surveillance videos
SooknunanOptical/radio transient classification with machine learning
ThompsonThe use of contour integrals for error estimation in Gauss quadrature
TorpeyHuman action recognition using recurrent neural networks
ToussaintA knowledge-centric approach to modelling dynamic customer segmentation in domestic load research
van der WaltAre the mysterious dryland "Fairy Circles" the result of microbial phytopathogenesis?
VosMachine learning in astronomy
WoodfordThe concurrent development of neural network based simulators and controllers in the evolutionary robotics process
ZangwaDevelopment of an algorithm for automatic segmentation of foliage images from uncontrolled environment
Tuesday September 12th Poster Presentations
First Name Surname Poster Name
AlhassanRadio sources classification with machine learning techniques
AmimaApplications of machine learning to identify differently expressed genes in breast cancer patients
AnguluHierarchical age estimation using facial features
AniyanDeep learning for radio astronomy
ArthurThermal hydraulics and transient analysis of a nuclear research reactor using CFD code
AyamiImproving facial recognition algorithms with an application to DUT security
BayanaGender classification based on feature fusion
BesterModel-based reinforcement learning with parameterised action spaces
BooyseRecurrent neural networks for asset failure prediction
BreytenbachOn the use of neural networks for classifying time series
ChibuyeLow-resource language speech recognition using cross-language phoneme mapping
ChingozhaStabilisability preserving abstractions of control systems
ChougradA deep learning framework for breast cancer screening
DaudDisability, brain control and ML
de WetAutomatic recognition of code-switched South African speech
du ToitA comparative evaluation and analysis of open-source part-of-speech taggers for under resourced official South African languages
DufourqDeep neural network architecture optimisation
Egbelowo-
El BouchefryAstrophysical machine learning software: an overview
GouayaFace recognition using HOSVD
GrunowFull 3+1 dimensional simulation of the Boltzmann equation in the context of heavy ion collisions
GueorguievASTCVS: a cognitive vision system for astronomical phenomenon classification
HasaniLearning and modeling analog behavior
HooperActing under partial information in virtual worlds
IkramUsing deep learning with community detection to solve the cold start problem in recommender systems
Jain-
JosiasIdentifying vein intersections in Tsetse fly wing images
KeivaniTracking moving objects in dynamic scenes
KekereIncremental learning for big data
KhozaA spiking neural network approach to natural language processing
KnowlesGMRT diffuse radio emission cluster survey
KohlakalaHuman ear recognition based on global features
Lai HongAdaptive knowledge injection for Monte Carlo tree search
LamboDeep reinforcement learning for games of perfect information
MakatiSemi-automated detectable life band
ManabeAfrican Solutions to Meet SDG 7: Affordable and Clean Energy
MarumoData analytic framework for radio astronomy: hierarchical artifact detection (HArD)
MasekoOptimized path planning and path tracking for autonomous vehicles with a constrained turning rate
Mbonda Tiekwe-
Modupe-
MoiloaSegmentation of low contrast time based neural fluorescence images
MokoenaAnomaly explanations
Moodley-
MostertCalibration of stochastic processes using neural networks
MotsoehliUse of spark streaming and gradient boosting to identify application fraud in the retail space in realtime
MpipiPredicting students performance using machine learning
MtetwaBrain MRI structure segementation
MuleyaMultiscale modeling TB transmission dynamics
MvubuFinancial market time series prediction with Long Short Term Memory (LSTM's)
NdlovuStereo image and 3D point cloud mapping and localisation for self-driving cars in outdoor unstructured environments
NewmanClassifying sleep stages from wearable data
NguyenLearning to rank from clicks: a probabilistic model for separating relevance from positional bias in search logs
NiitApplications of reinforcement learning to psychiatric disorders
NyambalAutomatic parking space detection based on convolutional neural networks and support vector machines
Ofosu MensahInvestigating relationship between expressed cancer related genes and patient survival
OkoumaA new deterministic scheme for characterizing the organization of (large) prime numbers
OldewageParticle swarm optimization in high dimensional spaces
PerezSupporting the fight against human trafficking
PretoriusLearning dynamics of neural networks in reinforcement learning
RadulescuWhole body locomotion planning for quadrupedal systems
RajaramGenerative adversarial networks
RanchodSkill discovery in reinforcement learning
RapheehaArtificial neural network
RiceDiscrete mathematics intelligent tutoring system using Bayesian networks
SchuldMachine learning on quantum computers
Sefala3D convolutions for action recognition
SihlanguCognitive MeerKat
SitholeBig data and loT
SteynShort-term stream flow forecasting at Australian river sites using data-driven regression techniques
TakongInvestigating rainfall spatial variability over the Drakensberg Mountain Range using artificial neural networks
TavakoliDeep reinforcement learning for robot learning
UmuhireDevelopment of a post-processing technique for a quantum key distribution system
van BiljonThe creation of intelligent agents with machine learning
van NiekerkDirect sampling using the discrete pulse transform
van NiekerkOnline constrained model-based reinforcement learning
ZithaFine-tuning radio calibration using machine learning methods
ZwaneRobot multi-tasking using deep reinforcement learning from raw RGBDT data (T for tactile sensor data)

Poster Preparation and Prizes

  • Poster size: Should be A0 size, in either landscape or portrait mode.
  • Poster printing: We suggest you print your poster in advance and bring it with you.
  • If you need inspiration in designing a poster, try looking here or here for some advice.

Prizes!

We want everyone to have fun at these events, but also to encourage people to do their best work. The best posters will receive prizes for their efforts. We have these prizes, and a few other secret prizes to be awarded on the final night to those posters and presenters that have shined.

NVIDIA
Cambridge University Press
MIT press online

If there is anything we can clarify, please reach out to us at: dl-indaba@googlegroups.com