South African Tuberculosis Vaccine Initiative
UCT Computer Science Research Document ArchiveNot a member yet
1270 research outputs found
Sort by
Evolving Herding Behaviour Diversity in Robot Swarms
Full text linkBehavioural diversity has been demonstrated as beneficial in biological
social systems, such as insect colonies and human societies,
as well as artificial systems such as large-scale software and
swarm-robotics systems. Evolutionary swarm robotics is a popular
experimental platform for demonstrating the emergence of
various social phenomena and collective behaviour, including behavioural
diversity and specialization. However, from an automated
design perspective, the evolutionary conditions necessary to synthesize
optimal collective behaviours (swarm-robotic controllers)
that function across increasingly complex environments (difficult
tasks), remains unclear. Thus, we introduce a comparative study
of behavioural-diversity maintenance methods (swarm-controller
extension of the MAP-Elites algorithm) versus those without behavioural
diversity mechanisms (Steady-State Genetic Algorithm),
as a means to evolve suitable degrees of behavioural diversity over
increasingly difficult collective behaviour (sheep-dog herding) tasks.
In support of previous work, experiment results demonstrate that
behavioural diversity can be generated without specific speciation
mechanisms or geographical isolation in the task environment, although
the direct evolution of a functionally (behaviorally) diverse
swarm does not yield high task performance
The Impact of Morphological Diversity in Robot Swarms
Full text linkIn nature, morphological diversity enhances functional diversity,
however, there is little swarm (collective) robotics research on the
impact of morphological and behavioral (body-brain) diversity that
emerges in response to changing environments. This study investigates
the impact of increasingly complex task environments on
the artificial evolution of body-brain diversity in simulated robot
swarms. We investigate whether increasing task environment complexity
(collective behavior tasks requiring increasing degrees of
cooperative behavior) mandates concurrent increases in behavioral,
morphological, or coupled increases in body-brain diversity
in robotic swarms. Experiments compared three variants of collective
behavior evolution across increasingly complex task environments:
two behavioral diversity maintenance variants and
body-brain diversity maintenance. Results indicate that body-brain
diversity maintenance yielded a significantly higher behavioral
and morphological diversity in evolved swarms overall, which was
beneficial in the most complex task environment
Conformational comparisons of Pasteurella multocida types B and E and structurally related capsular polysaccharides
Full text linkPasteurella multocida, an encapsulated gram-negative bacterium, is a significant veterinary pathogen. The P. multocida is classified into 5 serogroups (A, B, D, E, and F) based on the bacterial capsular polysaccharide (CPS), which is important for virulence. Serogroups B and E are the primary causative agents of bovine hemorrhagic septicemia that is associated with significant yearly losses of livestock worldwide, primarily in low- and middle-income countries. The P. multocida disease is currently managed by whole-cell vaccination, albeit with limited efficacy. CPS is an attractive antigen target for an improved vaccine: CPS-based vaccines have proven highly effective against human bacterial diseases and could provide longer-term protection against P. multocida. The recently elucidated CPS repeat units of serogroups B and E both comprise a N-acetyl-β-D-mannosaminuronic acid/N-acetyl-β-D-glucosamine disaccharide backbone with β-D-fructofuranose (Fruf) side chain, but differ in their glycosidic linkages, and a glycine (Gly) side chain in serogroup B. Interestingly, the Haemophilus influenzae types e and d CPS have the same backbone residues. Here, comparative modeling of P. multocida serogroups B and E and H. influenzae types e and d CPS identifies a significant impact of small structural differences on both the chain conformation and the exposed potential antibody-binding epitopes (Ep). Further, Fruf and/or Gly side chains shield the immunogenic amino-sugar CPS backbone—a possible common strategy for immune evasion in both P. multocida and H. influenzae. As the lack of common epitopes suggests limited potential for cross-reactivity, a bivalent CPS-based vaccine may be necessary to provide adequate protection against P. multocida types B and E
Subword Segmental Machine Translation: Unifying Segmentation and Target Sentence Generation
Full text linkSubword segmenters like BPE operate as a pre-processing step in neural machine translation
and other (conditional) language models. They are applied to datasets before training, so translation or text generation quality relies on the quality of segmentations. We propose a departure from this paradigm, called subword segmental machine translation (SSMT). SSMT
unifies subword segmentation and MT in a single trainable model. It learns to segment target sentence words while jointly learning to generate target sentences. To use SSMT during inference we propose dynamic decoding, a text generation algorithm that adapts segmentations as it generates translations. Experiments across 6 translation directions show that SSMT improves chrF scores for morphologically rich agglutinative languages. Gains are strongest in the very low-resource scenario. SSMT also learns subwords that are closer to morphemes compared to baselines and proves more robust on a test set constructed for evaluating morphological compositional generalisation
Reconsidering Network Management Interfaces for Communities
Full text linkCommunity-owned mesh wireless networks enable cost-effective sharing of networked resources, expanding internet and local service accessibility through low-cost WiFi hardware. However, maintaining these networks comes with expenses. In addition to hardware costs, community members need extensive training to install, monitor, and troubleshoot the networks using Network Management Interfaces (NMIs). Effective network management is crucial for CWN resilience within communities. This paper presents qualitative interviews with 25 stakeholders from two CWNs in India and four in South Africa, examining challenges to CWN resilience. Workshops were conducted with network operators and users in India (Janastu) and prospective operators in South Africa (FOCUS Network) to reimagine NMIs, discussing challenges and prototyping interfaces. Our findings highlights diverse network management approaches, revealing difficulties in technical capacity building, troubleshooting, and prototyping. Designing NMIs with local network operators’ insights and skills is crucial for CWN sustainability. The paper outlines design opportunities to improve network management interfaces for CWNs, fostering network resilience for critical infrastructures
Multilingual Knowledge Graphs and Low-Resource Languages: A Review
No full textThere is a lack of multilingual data to support applications in a large number of languages, especially for low-resource languages. Knowledge graphs (KG) could contribute to closing the gap of language support by providing easily accessible, machine-readable, multilingual linked data, which can be reused across applications. In this paper, we provide an overview of work in the domain of multilingual KGs with a focus on low-resource languages. We review the current state of multilingual KGs along with the different aspects that are crucial for creating KGs with language coverage in mind. Special consideration is given to challenges particular to low-resource languages in KGs. We further provide an overview of applications that yield multilingual KG information as well as downstream applications reusing such multilingual data. Finally, we explore open problems regarding multilingual KGs with a focus on low-resource languages
An analysis of positionalism's roles in use
Full text linkRoles as argument places in a positionalist account of relations are pervasive in conceptual data modelling and linguistics, known also as components of a relationship and as semantic, thematic, or deep roles as parts of a verb or verb class, respectively. They are also planned to be used in Abstract Wikipedia that seeks to combine them. There is, however, no insight in systematic or ontologically sound usage of such roles, in contradistinction to the ample attention given to aligning classes to nouns and relationships to verbs. Roles, as identifiable argument places, may benefit from similar efforts toward an ontology of roles. We aim to take a first step in that direction in a two-pronged approach. First, we conducted an analysis of a set of 101 conceptual data models on their use of roles. Second, we analysed VerbNet, an authoritative linguistic knowledge base on thematic roles. The results show promise for improvements of naming roles in conceptual data models. VerbNet’s roles are challenging to align to an ontology due to its mixing of the ontological and linguistic layers and flexibility of natural language. The insights obtained also indicate ample avenues for further research
A Framework for Interoperability Between Models with Hybrid Tools
Full text linkComplex system development and maintenance face the challenge of dealing with different
types of models due to language affordances, preferences, sizes, and so forth that involve
interaction between users with different levels of proficiency. Current conceptual data
modelling tools do not fully support these modes of working. It requires that the interaction between multiple models in multiple languages is clearly specified to ensure they keep their intended semantics, which is lacking in extant tools. The key objective is to devise a mechanism to support semantic interoperability in hybrid tools for multi-modal modelling in a plurality of paradigms, all within one system. We propose FaCIL, a framework for such hybrid modelling tools. We design and realise the framework FaCIL, which maps UML, ER and ORM2 into a common metamodel with rules that provide the central point for management among the models and that links to the formalisation and logic-based
automated reasoning. FaCIL supports the ability to represent models in different formats
while preserving their semantics, and several editing workflows are supported within the
framework. It has a clear separation of concerns for typical conceptual modelling activities in an interoperable and extensible way. FaCIL structures and facilitates the interaction between visual and textual conceptual models, their formal specifications, and abstractions as well as tracking and propagating updates across all the representations. FaCIL is compared against the requirements, implemented in crowd 2.0, and assessed with a use case. The proof-of-concept implementation in the web-based modelling tool crowd 2.0 demonstrates its viability. The framework also meets the requirements and fully supports the use case
Using Graph Theory to Produce Emergent Behaviour in Agent-Based Systems
Full text linkCooperation is a defining trait of Multi-Agent Systems.
At the centre of these systems lies a communication network
which governs how information flows from one agent to the next.
However, the design of these networks is often overlooked despite
the profound impact it can have on both the task performance
of the agents and the emergent phenomena they produce. In this
work we aim to illustrate this by investigating whether network
centrality impacts the task performance and emergent inequality
(unequal distribution of resources) of resource gathering agents.
We achieve this by constructing several communication networks
with increasing centrality and use them with an Agent-Based
Model called GATHER. Our results indicate that as the variance
of the population’s centrality increases, the task performance of
an agent population will decrease. Furthermore, we demonstrate
that simply changing the centrality of the network can produce
distinct results and emergent phenomena (inequality or the lack
thereof in our case). We then further support this claim by
increasing the reciprocity of one of our communication networks
which results in a system with greater task performance and
significantly lower inequality, further illustrating the impact communication
network topology can have on Multi-Agent Systems
CoLRN - A Community-Based Vision for Local Resilient Networks
Full text linkIn this research, we share our findings from a series of design workshops with community wireless network members and their users in India and Africa to develop a community-based vision for resilient local networks. We simultaneously leveraged existing projects in India and South Africa around network management interfaces and local content creation to evaluate our design strategies to foster resilience and effectiveness in empowering community networks. Through this work, we identified the challenges and opportunities for innovative approaches to leveraging networked technologies to bring communities together to learn from each other on how they manage and use their community network. We highlight key opportunities to explore a) infrastructural resilience through community-centred design of network management tools, and b) novel approaches to support content creation tapping community desires to capture local knowledge, through annotation of digital stories and production of radio content