NRU :: Browsing by Author "Bainomugisha, Engineer"

Browsing by Author "Bainomugisha, Engineer"

Now showing 1 - 20 of 36

Air pollution and mobility patterns in two Ugandan cities during COVID‑19 mobility restrictions suggest the validity of air quality data as a measure for human mobility
(Environmental Science and Pollution Research, 2022) Galiwango, Ronald; Bainomugisha, Engineer; Kivunike, Florence; Kateete, David Patrick; Jjingo, Daudi
We explored the viability of using air quality as an alternative to aggregated location data from mobile phones in the two most populated cities in Uganda. We accessed air quality and Google mobility data collected from 15th February 2020 to 10th June 2021 and augmented them with mobility restrictions implemented during the COVID-19 lockdown. We determined whether air quality data depicted similar patterns to mobility data before, during, and after the lockdown and determined associations between air quality and mobility by computing Pearson correlation coefficients ( R ), conducting multivariable regression with associated confidence intervals (CIs), and visualized the relationships using scatter plots. Residential mobility increased with the stringency of restrictions while both non-residential mobility and air pollution decreased with the stringency of restrictions. In Kampala, PM2.5 was positively correlated with non-residential mobility and negatively correlated with residential mobility. Only correlations between PM2.5 and movement in work and residential places were statistically significant in Wakiso. After controlling for stringency in restrictions, air quality in Kampala was independently correlated with movement in retail and recreation (− 0.55; 95% CI = − 1.01– − 0.10), parks (0.29; 95% CI = 0.03–0.54), transit stations (0.29; 95% CI = 0.16–0.42), work (− 0.25; 95% CI = − 0.43– − 0.08), and residential places (− 1.02; 95% CI = − 1.4– − 0.64). For Wakiso, only the correlation between air quality and residential mobility was statistically significant (− 0.99; 95% CI = − 1.34– − 0.65). These findings suggest that air quality is linked to mobility and thus could be used by public health programs in monitoring movement patterns and the spread of infectious diseases without compromising on individuals’ privacy.
Air quality management strategies in Africa: A scoping review of the content, context, co-benefits and unintended consequences
(Environment International, 2022) Okello, Gabriel; Nantanda, Rebecca; Awokola, Babatunde; Thondoo, Meelan; Okure, Deo; Tatah, Lambed; Bainomugisha, Engineer; Oni, Tolu
One of the major consequences of Africa’s rapid urbanisation is the worsening air pollution, especially in urban centres. However, existing societal challenges such as recovery from the COVID-19 pandemic, poverty, intensifying effects of climate change are making prioritisation of addressing air pollution harder. We undertook a scoping review of strategies developed and/or implemented in Africa to provide a repository to stakeholders as a reference that could be applied for various local contexts. The review includes strategies assessed for effectiveness in improving air quality and/or health outcomes, co-benefits of the strategies, potential collaborators, and pitfalls. An international multidisciplinary team convened to develop well-considered research themes and scope from a contextual lens relevant to the African continent. From the initial 18,684 search returns, additional 43 returns through reference chaining, contacting topic experts and policy makers, 65 studies and reports were included for final analysis. Three main strategy categories obtained from the review included technology (75%), policy (20%) and education/ behavioural change (5%). Most strategies (83%) predominantly focused on household air pollution compared to outdoor air pollution (17%) yet the latter is increasing due to urbanisation. Mobility strategies were only 6% compared to household energy strategies (88%) yet motorised mobility has rapidly increased over recent decades. A cost effective way to tackle air pollution in African cities given the competing priorities could be by leveraging and adopting implemented strategies, collaborating with actors involved whilst considering local contextual factors. Lessons and best practices from early adopters/implementers can go a long way in identifying opportunities and mitigating potential barriers related to the air quality management strategies hence saving time on trying to “reinvent the wheel” and prevent pitfalls. We suggest collaboration of various stakeholders, such as policy makers, academia, businesses and communities in order to formulate strategies that are suitable and practical to various local contexts.
Applying machine learning for large scale field calibration of low-cost PM2.5 and PM10 air pollution sensors
(Applied AI Letters, 2022) Adong, Priscilla; Bainomugisha, Engineer; Okure, Deo; Sserunjogi, Richard
Low-cost air quality monitoring networks can potentially increase the availability of high-resolution monitoring to inform analytic and evidence-informed approaches to better manage air quality. This is particularly relevant in low and middle-income settings where access to traditional reference-grade monitoring networks remains a challenge. However, low-cost air quality sensors are impacted by ambient conditions which could lead to over- or underestimation of pollution concentrations and thus require field calibration to improve their accuracy and reliability. In this paper, we demonstrate the feasibility of using machine learning methods for large-scale calibration of AirQo sensors, lowcost PM sensors custom-designed for and deployed in Sub-Saharan urban settings. The performance of various machine learning methods is assessed by comparing model corrected PM using k-nearest neighbours, support vector regression, multivariate linear regression, ridge regression, lasso regression, elastic net regression, XGBoost, multilayer perceptron, random forest and gradient boosting with collocated reference PM concentrations from a Beta Attenuation Monitor (BAM). To this end, random forest and lasso regression models were superior for PM2.5 and PM10 calibration, respectively. Employing the random forest model decreased RMSE of raw data from 18.6 μg/m3 to 7.2 μg/m3 with an average BAM PM2.5 concentration of 37.8 μg/m3 while the lasso regression model decreased RMSE from 13.4 μg/m3 to 7.9 μg/m3 with an average BAM PM10 concentration of 51.1 μg/m3. We validate our models through cross-unit and cross-site validation, allowing analysis of AirQo devices' consistency. The resulting calibration models were deployed to the entire large-scale air quality monitoring network consisting of over 120 AirQo devices, which demonstrates the use of machine learning systems to address practical challenges in a developing world setting.
Bringing Scheme Programming to the iPhone - Experience
(Software: Practice and Experience, 2012) Bainomugisha, Engineer; Vallejos, Jorge; Gonzalez Boix, Elisa; Costanza, Pascal; D'Hondt, Theo; De Meuter, Wolfgang
The iPhone SDK provides a powerful platform for the development of applications that make use of iPhone capabilities such as sensors, GPS, Wi-Fi or Bluetooth connectivity. Thus far we observe that the development of iPhone applications is mostly restricted to using Objective-C. However, developing applications in plain Objective-C on the iPhone OS su ers from limitations such as the need for explicit memory management and lack of syntactic extension mechanism. Moreover, when developing distributed applications in Objective-C, programmers have to manually deal with distribution concerns such as service discovery, remote communication, and failure handling. In this paper, we discuss our experience on porting the Scheme programming language to the iPhone OS and how it can be used together with Objective-C to develop iPhone applications. To support the interaction between Scheme programs and the underlying iPhone APIs, we have implemented a language symbiosis layer that enables programmers to access the iPhone SDK libraries from Scheme. In addition, we have designed high-level distribution constructs to ease the development of distributed iPhone applications in an event-driven style. We validate and discuss these constructs with a series of examples including an iPod controller, a maps application and a distributed multiplayer Scrabble-like game. We discuss the lessons learned from this experience for other programming language ports to mobile platforms.
Characterization of Ambient Air Quality in Selected Urban Areas in Uganda A Low-Cost Approach
(Environmental Science & Technology, 2022) Okure, Deo; Bainomugisha, Engineer; Lozano-Gracia, Nancy; Soppelsa, Maria Edisa
Many cities and urban centers around the world experience high air pollution episodes attributable to increased anthropogenic alterations of natural environmental systems. World Health Organization estimates indicate strong exceedances of prescribed limits in developing countries. However, the evidence on local pollution measures is limited for such cities and Uganda is no exception. Informed by the practical realities of air quality monitoring, this paper employs a low-cost approach using passive and active monitors to obtain characterization of pollution levels based on particulate matter 2.5, nitrogen dioxide, and ozone over a six-month period (starting in December 2018) for selected urban centers in three of the four macro-regions in Uganda. This is the first attempt to comprehensively assess pollution levels at a near-national level in Uganda. A combination of distributed stationary monitors and mobile monitors installed on motorcycle taxis (boda-boda) was employed in selected parishes to obtain spatiotemporal variations in the pollutant concentrations. The results suggest that seasonal particulate levels heavily depend on precipitation patterns with a strong inverse relation, which further corroborates the need for longer monitoring periods to reflect actual seasonal variations. Informed by the observed level of data completeness and quality in all the monitoring scenarios, the paper highlights the practicability and potential of a low-cost approach to air quality monitoring and the potential to use this information to inform citizens.
Clone-Based Variability Management in the Android Ecosystem
(IEEE, 2018) Businge, John; Openja, Moses; Nadi, Sarah; Bainomugisha, Engineer; Berger, Thorsten
Abstract—Mobile app developers often need to create variants to account for different customer segments, payment models or functionalities. A common strategy is to clone (or fork) an existing app and then adapt it to new requirements. This form of reuse has been enhanced with the advent of social-coding platforms such as Github, cultivating a more systematic reuse. Different facilities, such as forks, pull requests, and cross-project traceability support clone-based development. Unfortunately, even though, many apps are known to be maintained in many variants, little is known about how practitioners manage variants of mobile apps. We present a study that explores clone-based reuse practices for open-source Android apps. We identified and analyzed families of apps that are maintained together and that exist both on the official app store (Google Play) as well as on Github, allowing us to analyze reuse practices in depth. We mined both repositories to identify app families and to study their characteristics, including their variabilities as well as codepropagation practices and maintainer relationships. We found that, indeed, app families exist and that forked app variants fall into the following categories: (i) re-branding and simple customizations, (ii) feature extension, (iii) supporting of the mainline app, and (iv) implementation of different, but related features. Other notable characteristic of the app families we discovered include: (i) 73% of the app families did not perform any form of code propagation, and (ii) 74% of the app families we studied do not have common maintainers.
Code Authorship and Fault-proneness of Open-Source Android Applications : An Empirical Study
(International Conference on Predictive Models and Data Analytics in Software Engineering, 2017) Businge, John; Kawuma, Simon; Bainomugisha, Engineer; Khomh, Foutse; Nabaasa, Evarist
In recent years, many research studies have shown how human factors play a significant role in the quality of software components. Code authorship metrics have been introduced to establish a chain of responsibility and simplify management when assigning tasks in large and distributed software development teams. Researchers have investigated the relationship between code authorship metrics and fault occurrences in software systems. However, we have observed that these studies have only been carried on large software systems having hundreds to thousands of contributors. In our preliminary investigations on Android applications that are considered to be relatively small, we observed that applications systems are not totally owned by a single developer (as one could expect) and that cases of no clear authorship also exist like in large systems. To this end, we do believe that the Android applications could face the same challenges faced by large software systems and could also benefit from such studies. Goal: We investigate the extent to which the findings obtained on large software systems applies to Android applications. Approach: Building on the designs of previous studies, we analyze 278 Android applications carefully selected from GitHub. We extract code authorship metrics from the applications and examine the relationship between code authorship metrics and faults using statistical modeling. Results: Our analyses confirm most of the previous findings, i.e., Android applications with higher levels of code authorship among contributors experience fewer faults.
Context-oriented Programming for Customizable SaaS Applications
(Annual acm symposium on applied computing, 2012) Truyen, Eddy; Cardozo, Nicolás; Walraven, Stefan; Vallejos, Jorge; Bainomugisha, Engineer; Günther, Sebastian; D’Hondt, Theo; Joosen, Wouter
Software-as-a-Service (SaaS) applications are multi-tenant software applications that are delivered as highly configurable web services to individual customers, which are called tenants in this context. For reasons of complexity management and to lower maintenance cost, SaaS providers maintain and deploy a single version of the application code for all tenants. As a result, however, custom-made extensions for individual tenants cannot be efficiently integrated and managed. In this paper we show that by using a context-oriented programming model, cross-tier tenant-specific software variations can be easily integrated into the single-version application code base. Moreover, the selection of which variations to execute can be configured on a per tenant basis. Concretely, we provide a technical case study based on Google App Engine (GAE), a cloud platform for building multitenant web applications. We contribute by showing: (a) how ContextJ, a context-oriented programming (COP) language, can be used with GAE, (b) the increase in flexibility and customizability of tenant-specific software variations using ContextJ as compared to Google’s dependency injection framework Guice, and (c) that the performance of using ContextJ is comparable to Guice. Based on these observations, we come to the conclusion that COP can be helpful for providing software variations in SaaS.
Crane Cloud: A resilient multi-cloud service abstraction layer for resource-constrained settings
(Development Engineering, 2022) Bainomugisha, Engineer; Mwotil, Alex
Developers and users situated in low-resource settings are faced with unique contextual and infrastructure challenges when accessing and consuming cloud-based services. In low-resource settings, access to cloud services and platforms is usually characterized by low-end computing devices and often unreliable and slow mobile broadband Internet connections. In this paper, we discuss key challenges for developing for and accessing cloud services in resource constrained settings, namely, (1) Frequent Internet partitions and bandwidth constraints, (2) Data jurisdiction restrictions, (3) Vendor lock-in, and (4) Poor quality of service. Inspired by these challenges, we propose a set of important design considerations and properties for a resilient multi-cloud service layer, that includes: (1) Containerization and orchestration of applications, (2) Application placement and replication, (3) Portability and multi-cloud migration, (4) Resilience to network partitions and bandwidth constraints, (5) Automated service discovery and load balancing, (6) Localized image registry, and (7) Support for platform monitoring and management. We present an implementation and validation case study, Crane Cloud, an open source multi-cloud service abstraction layer built on-top of Kubernetes that is designed with inherent support for resilience to network partitions, microservice orchestration (deployment, scaling and management of containerized applications), a localized image registry, support for migration of services between private and public clouds to avoid vendor lock-in issues and platform monitoring. We evaluate the performance and user experience of Crane Cloud by implementing and deploying a computational and bandwidth intensive machine learning system. The results show lower response times of the system on Crane Cloud compared with hosting on other public clouds. The Crane Cloud platform is serving as a cloud-service for students and developers in low-resource settings and also as an education platform for cloud computing.
Crane Cloud: a resilient multi-cloud service layer for resource constrained settings
(Makerere University, 2021) Bainomugisha, Engineer; Mwotil, Alex
Whereas the main cloud providers have set up cloud services on stable infrastructure, developers and users situated in low-resource settings access cloud services and platforms using low-end computing devices that often connect to the Internet via slow mobile connections. These settings require custom software abstraction layers that consider such bandwidth constraints and intermittent connections as a rule rather than the exception. In this paper, we identify key challenges for developing for and accessing cloud services in resource constrained settings, namely, (1) Frequent Internet partitions and bandwidth constraints, (2) Data jurisdiction restrictions, (3) Vendor lock-in, and (4) Poor quality of service. To address these challenges, we propose a set of design considerations and properties for a resilient multi-cloud service layer, that includes: (1) Containerisation and orchestration of applications, (2) Service scheduling and replication, (3) Portability and multi-cloud migration, (4) Resilience to network partitions and bandwidth constraints, (5) Automated service discovery and load balancing, (6) Localised image registry, and (7) Support for platform monitoring and management. We present a prototype validation case study, Crane Cloud, an open source multi-cloud service abstraction layer built on-top of Kubernetes that is designed with inherent support for resilience to network partitions, microservice orchestration (deployment, scaling and management of containerized applications)a localized image registry, support for migration of services between private and public clouds to avoid vendor lock-in issues and platform monitoring. We evaluate the performance and user experience of Crane Cloud by implementing and deploying a computational and bandwidth intensive machine learning system shows lower response time compared when hosted on other public clouds.
Data Classification for Secure Mobile Health Data Collection Systems
(Development Engineering, 2020) Katarahweire, Marriette; Bainomugisha, Engineer; Mughal, Khalid A.
Data collected in Mobile Health Data Collections Systems (MHDCS) are diverse, both in terms of type and value. This calls for different data protection measures to meet security goals of confidentiality, integrity, and availability. The majority of commonly used open-source MHDCS track and monitor individuals over a while. It is therefore important to have sensitive data defined and proper security measures identified. We propose a data classification model as a basis for secure design and implementation. Our method combines interviews with case studies. The case studies focused on three of the widely used MHDCS platforms in low-resource settings; that is Muzima, Open Data Kit (ODK), and District Health Information Software (DHIS) 2 Tracker Capture. Interviews with domain experts helped define the sensitivity of data in MHDCS. The proposed data classification model provides for three sensitivity levels: public, confidential, and critical. The model uses context information and multiple parameters as inputs to a classification scheme that maps data to sensitivity levels. The generated data classifications are intended to guide developers and users to build security into MHDCS starting from the early stages of the software development life cycle.
Determination of Satellite-Derived PM2.5 for Kampala District, Uganda
(Geomatics, 2022) Atuhaire, Christine; Gidudu, Anthony; Bainomugisha, Engineer; Mazimwe, Allan
Ground monitoring stations are widely used to monitor particulate matter (PM2.5). However, they are expensive to maintain and provide information localized to the stations, and hence are limited for large-scale use. Analysis of in situ PM2.5 shows that it varies spatially and temporally with distinct seasonal differences. This study, therefore, explored the use of satellite images (Sentinel-2 and Landsat-8) for determining the spatial and temporal variations in PM2.5 for Kampala District in Uganda. Firstly, satellite-derived aerosol optical depth (AOD) was computed using the Code for High Resolution Satellite mapping of optical Thickness and aNgstrom Exponent algorithm (CHRISTINE code). The derived AOD was then characterised with reference to meteorological factors and then correlated with in situ PM2.5 to determine satellite-derived PM2.5 using geographically weighted regression. In the results, correlating in situ PM2.5 and AOD revealed that the relationship is highly variable over time and thus needs to be modelled for each satellite’s overpass time, rather than having a generic model fitting, say, a season. The satellite-derived PM2.5 showed good model performance with coefficient of correlation (R2) values from 0.69 to 0.89. Furthermore, Sentinel-2 data produced better predictions, signifying that increasing the spatial resolution can improve satellite-derived PM2.5 estimations.
Exploring PM2.5 variations from calibrated low-cost sensor network in Greater Kampala, during COVID-19 imposed lockdown restrictions: Lessons for Policy
(Clean Air Journal, 2022) Green, Paul; Okure, Deo; Adong, Priscilla; Sserunjogi, Richard; Bainomugisha, Engineer
Air pollution is considered a major public health risk globally, and the global South including sub-Saharan Africa face particular health risks, but there is limited data to quantify the level of pollution for different air quality contexts. The COVID-19 lockdown measures led to reduced human activities, and provided a unique opportunity to explore the impacts of reduced activities on urban air quality. This paper utilises calibrated data from a low-cost sensor network to explore insights from the diverse ambient air quality profile for four urban locations in Greater Kampala, Uganda before and during lockdown from March 31 to May 5 2020, highlighting the uniqueness of air pollution profiles in a sub-Saran Africa context. All locations saw year to year improvements in 24-hour mean PM2.5 between 9 and 25μg/m3 (i.e. 17-50% reduction from the previous year) and correlated well with reduction in traffic (up to approx. 80%) and commercial activities. The greatest improvement was observed in locations close to major transport routes in densely populated residential areas between 8 pm and 5 am. This suggests that the reduction in localised pollution sources such as nocturnal polluting activities including traffic and outdoor combustion including street cooking characteristic of fast-growing cities in developing countries, coupled with meteorological effects led to amplified reductions that continued well into the night, although meteorological effects are more generalised. Blanket policy initiatives targeting peak pollution hours could be adopted across all locations, while transport sector regulation could be very effective for pollution management. Likewise, because of the clustered and diffuse nature of pollution, community driven initiatives could be feasible for long-term mitigation.
Flexub: Dynamic Subscriptions for Publish/Subscribe Systems in MANETs
(Springer Berlin Heidelberg, 2012) Bainomugisha, Engineer; Paridel, Koosha; Vallejos, Jorge; Berbers, Yolande; De Meuter, Wolfgang
Current publish/subscribe systems provide very limited support to modify subscriptions dynamically. Consequently, they cannot efficiently control the flow of events between publishers and subscribers, which may lead to unnecessary network traffic. In addition, it is not possible to automatically subscribe or unsubscribe to a service depending on certain context of use. This implies for developers to manually manage subscriptions (e.g., taking care of when to cancel or re-issue a subscription), which may result in inappropriate subscription states (e.g., subscriptions that are cancelled too late). In this paper, we propose the concept of dynamic subscription mechanisms that improves the expressiveness and flexibility of subscriptions. We introduce a new dimension to a subscription that allows a subscriber to express the flow of matched events, and when a new subscription can be (re)issued. We validate our claims for improved flexibility and expressiveness by providing language abstractions and a prototype implementation of a dynamic subscription mechanism framework called Flexub that supports a variation of subscription mechanisms. When compared to existing subscription models, our experiment results show that the support for dynamic subscription mechanisms greatly reduces network traffic of events sent from publishers to the subscribers. In addition, our approach reduces the workload on the subscriber side.
Gaussian Process Models for Low Cost Air Quality Monitoring
(University of Makerere, 2021) Smith, Michael T.; Ssematimba, Joel; Alvarez, Mauricio A.; Bainomugisha, Engineer
Air pollution contributes to over three million deaths [1] each year. Kampala has one of the highest concentrations of fine particulate matter (PM 2.5) of any African city [2]. Unfortunately, with the exception of the US Embassy, there is no programme for monitoring air pollution in the city due to the high cost of the equipment required. Hence we know little about its distribution or extent. Lower cost devices do exist, but these do not, on their own, provide the accuracy required for decision makers. We propose that using a coregionalised Gaussian process to combine the low cost sensors with the embassy’s high quality results provides sufficiently accurate estimates of pollution across the city.
How Stable are Eclipse Application Framework Internal Interfaces?
(IEEE, 2019) Businge, John; Kawuma, Simon; Openja, Moses; Bainomugisha, Engineer; Serebrenik, lexander
Eclipse framework provides two interfaces: stable interfaces (APIs) and unstable interfaces (non-APIs). Despite the non-APIs being discouraged and unsupported, their usage is not uncommon. Previous studies showed that applications using relatively old non-APIs are more likely to be compatible with new releases compared to the ones that used newly introduced non-APIs; that the growth rate of non-APIs is nearly twice as much as that of APIs; and that the promotion of non-API to APIs happens at a slow pace since API providers have no assistance to identify public interface candidates. Motivated by these findings, our main aim was to empirically investigate the entire population (2,380K) of non-APIs to find the non-APIs that remain stable for a long period of time. We employ cross-project clone detection to identify whether non- APIs introduced in a given Eclipse release remain stable over successive releases. We provide a dataset of 327K stable non- API methods that can be used by both Eclipse interface providers as possible candidates of promotion. Instead of promoting non- APIs which are too fine-grained, we summarized the non-API methods groups in given classes that are stable together and present class-level non-APIs that possible candidates promotion. We have shown that it is possible to predict the stability of a non-API in subsequent Eclipse releases with a precision of 56%, a recall of 96% and an AUC of 92% and an Fmeasure of 81%. We have also shown that the metrics of length of a method and number of method parameters in a non- API method are very good predictors for the stability of the non-API in successive Eclipse releases. The results provided can help the API providers to estimate a priori how much work could be involved in performing the promotion
How Stable are Eclipse Application Framework Internal Interfaces?
(IEEE, 2019) Businge, John; Kawuma, Simon; Openja, Moses; Bainomugisha, Engineer; Serebrenik, Alexander
Eclipse framework provides two interfaces: stable interfaces (APIs) and unstable interfaces (non-APIs). Despite the non-APIs being discouraged and unsupported, their usage is not uncommon. Previous studies showed that applications using relatively old non-APIs are more likely to be compatible with new releases compared to the ones that used newly introduced non-APIs; that the growth rate of non-APIs is nearly twice as much as that of APIs; and that the promotion of non-API to APIs happens at a slow pace since API providers have no assistance to identify public interface candidates. Motivated by these findings, our main aim was to empirically investigate the entire population (2,380K) of non-APIs to find the non-APIs that remain stable for a long period of time. We employ cross-project clone detection to identify whether non- APIs introduced in a given Eclipse release remain stable over successive releases. We provide a dataset of 327K stable non- API methods that can be used by both Eclipse interface providers as possible candidates of promotion. Instead of promoting non- APIs which are too fine-grained, we summarized the non-API methods groups in given classes that are stable together and present class-level non-APIs that possible candidates promotion. We have shown that it is possible to predict the stability of a non-API in subsequent Eclipse releases with a precision of 56%, a recall of 96% and an AUC of 92% and an Fmeasure of 81%. We have also shown that the metrics of length of a method and number of method parameters in a non- API method are very good predictors for the stability of the non-API in successive Eclipse releases. The results provided can help the API providers to estimate a priori how much work could be involved in performing the promotion.
Machine Learning for a Low-cost Air Pollution Network
(arXiv preprint, 2019) Smith, Michael T.; Ssematimba, Joel; Álvarez, Mauricio A.; Bainomugisha, Engineer
Data collection in economically constrained countries often necessitates using approximate and biased measurements due to the low-cost of the sensors used. This leads to potentially invalid predictions and poor policies or decision making. This is especially an issue if methods from resource-rich regions are applied without handling these additional constraints. In this paper we show, through the use of an air pollution network example, how using probabilistic machine learning can mitigate some of the technical constraints. Specifically we experiment with modelling the calibration for individual sensors as either distributions or Gaussian processes over time, and discuss the wider issues around the decision process.
Machine Translation for African Languages: Community Creation of Datasets and Models in Uganda
(n African Natural Language Processing, 2022) Akera, Benjamin; Mukiibi, Jonathan; Sanyu Naggayi, Lydia; Babirye, Claire; Owomugisha, Isaac; Nsumba, Solomon; Nakatumba-Nabende, Joyce; Bainomugisha, Engineer; Mwebaze, Ernest; Quinn, John
Reliable machine translation systems are only available for a small proportion of the world’s languages, the key limitation being a shortage of training and evaluation data. We provide a case study in the creation of such resources by NLP teams who are local to the communities in which these languages are spoken. A parallel text corpus, SALT, was created for five Ugandan languages (Luganda, Runyankole, Acholi, Lugbara and Ateso) and various methods were explored to train and evaluate translation models. The resulting models were found to be effective for practical translation applications, even for those languages with no previous NLP data available, achieving mean BLEU score of 26.2 for translations to English, and 19.9 from English. The SALT dataset and models described are publicly available at
Middleware for the Internet of Things, Design Goals and Challenges
(Electronic Communications of the EASST, 2010) Paridel, Koosha; Bainomugisha, Engineer; Vanrompay, Yves; Berbers, Yolande; De Meuter, Wolfgang
As the number of wireless devices increases and their size becomes smaller, there can be more interaction between everyday objects of our life. With advances in RFID chips and the introduction of new generations of these devices that are smaller and cheaper, it is possible to put a wireless interface on almost all everyday objects: vehicles, clothes, foodstuffs, etc. This concept is called the Internet of Things. Interaction with thousands of wireless devices leads to a continuous and massive flow of events which are generated spontaneously. The question of how to deal with this enormous number of events is challenging and introduces new design goals for a communication mechanism. In this paper we argue that a middleware together with suitable linguistic abstractions is a proper solution. We also point out the challenges in developing this middleware. Moreover, we give an overview of recent related work and describe why they fail to address these challenges.

Browsing by Author "Bainomugisha, Engineer"

Results Per Page

Sort Options