Theses - Electronics, Computer & Software Engineering TUS:MM
https://research.thea.ie/handle/20.500.12065/2472
2024-03-28T21:02:02ZAdaptive intent realisation (AIR) - inductive intent realisation through NLOP enabled intent matching
https://research.thea.ie/handle/20.500.12065/4677
Adaptive intent realisation (AIR) - inductive intent realisation through NLOP enabled intent matching
McNamara, Joseph
There is a strong interest in designing systems that can simplify the interactions between
humans and complex digital systems. Network Operators want more straightforward
mechanisms to engage with their networks and inform their actions and goals. Intent was
proposed to meet this challenge, but comes at a cost. Intent introduces large modelling
efforts, requiring Network Operators to gain expertise in formal model notation and the
integration of these models with their network. The cost is compounded by the speed
which modern networks evolve, requiring constant adaption to maintain intent-driven
features.
This work aims to leverage the concepts of intent-based management for private net works, without component and formal model expertise. This will be achieved through
the coordination of three enablers, Adaptive Policy, Machine Learning and Intent. Adap tive Policy provides a flexible framework for context-aware decision making, utilising a
state-based approach to policy execution. Machine Learning informs the decision mak ing process to produce impact-aware responses based on closed-loop reporting. Intent
structures the realisation process, how abstraction is handled through inductive pro cesses to generate actionable output. This work is highly experimental, developed on
site at the Network Management Lab in an Ericsson Product Development Unit based in
Ireland. This work concludes with the Adaptive Intent Realisation (AIR) reference ar chitecture successfully demonstrated in three use cases hosted in industrial grade private
5G networks.
2023-09-01T00:00:00ZReal-time processing of I-LOFAR data using signal and image-based artificial intelligence/machine learning methods
https://research.thea.ie/handle/20.500.12065/4676
Real-time processing of I-LOFAR data using signal and image-based artificial intelligence/machine learning methods
Scully, Jeremiah
Solar flares discharge up to 1025J of magnetic energy into the solar atmosphere and
are often linked with high-intensity radio emissions known as Solar Radio Bursts
(SRBs). SRBs are commonly found in dynamic spectra and are classified into five
major spectral classes, ranging from Type I to Type V, based on their form and
frequency, and time extent. The automatic detection and classification of such radio
bursts is a challenge in solar radio physics due to their heterogeneous form. Near
real-time detection and classification of SRBs has become a necessity in recent
years due to large data rates generated by advanced radio telescopes such as the
LOw-FRequency ARray (LOFAR).
This thesis proposes a strategy for developing a very fast image and signal classification system that uses artificial intelligence algorithms to process gigabyte/sec
data streams in real-time using the Irish-LOFAR array as its prime data source.
Currently, the state-of-the-art systems in this area are falling short of the required
performance to process such high-bandwidth data streams. Real-time study of SRBs
is crucial for effective solar monitoring because it provides timely information about
dynamic solar phenomena, such as flares and coronal mass ejections, allowing us
to predict space weather impacts on Earth’s technology and infrastructure. This
real-time data helps spacecraft operators, scientists, and public safety officials make
informed decisions, validates and refines models of solar behavior, and drives advancements in monitoring technologies, ensuring accurate and proactive responses
to solar disturbances
2023-08-01T00:00:00ZHead- and eye-based features for continuous core affect prediction
https://research.thea.ie/handle/20.500.12065/4018
Head- and eye-based features for continuous core affect prediction
O'Dwyer, Jonathan (Jonny)
Feelings, or affect, are a fundamental part of human experience. Arousal and valencemake up core affect and have received intense study in affective computing. Speechand facial features have been extensively studied as predictors of core affect. Otherindicators of affect include head- and eye-based gestures, yet these are underexploredfor affect prediction. In this dissertation, handcrafted feature sets from head and eyemodalities are proposed and evaluated in two audiovisual continuous (core) affectprediction experiments on the RECOLA and SEMAINE affective corpora.In the first experiment, head- and eye-based features were input to deep feed-forward neural network (DNN), along with speech and face features, for unimodalcontinuous affect prediction. Two proposed head feature sets and one eye featureset outperformed minimum performance benchmarks, estimated human predictionperformances, for arousal prediction on both corpora. The more complex headfeature set proposed performed second-best overall, after speech, and best from thevisual modalities, for arousal prediction. This feature set obtained validation setconcordance correlation coefficient (CCC) scores of 0.572 on RECOLA and 0.671on SEMAINE. For valence, head feature sets performed best from those proposed,and best overall for valence prediction on SEMAINE (CCC = 0.289), however,these sets were unable to match or exceed human performance estimates. From thisexperiment, it was concluded that head-based features are suitable for unimodalarousal prediction. It was also concluded that arousal prediction performance within-15.82% of speech, relative CCC, can be obtained from head-based features.In the second experiment, the proposed feature sets were evaluated with speechand face features for multimodal continuous affect prediction using DNNs. The ex-perimentation included a fusion study, cross-modal interaction feature investigation,and the proposal for, and evaluation of, teacher-forced learning with multi-stage re-gression (TFL-MSR). TFL-MSR is a method for leveraging correlations betweenaffect dimensions to improve affect prediction. An algorithm screening-based sensit-ivity was also performed to highlight important feature groups for prediction in thedifferent corpora. Model fusion performed better than feature fusion in the experi-ment. Relative CCC performance increases of 4.91% and 18.23% on RECOLA and13.18% and 74.17% on SEMAINE above model fusion speech and face were observedfor arousal and valence respectively for multimodal systems that used all modalit-ies. One eye and face cross-modal interaction feature was discovered for valenceprediction on RECOLA and it was able to improve CCC prediction performance by2.66%. TFL-MSR improved valence prediction on RECOLA but not on SEMAINEwhere a small arousal and valence correlation relationship was present. Interestingcross-corpus similarities and differences were found in the sensitivity analysis thatindicated some feature groups have similar importances, while other feature groups’importances were inverted across the social situations in the corpora. The finalmodels of this work produced test set CCC results of 0.812 for arousal and 0.463 forvalence on RECOLA and 0.616 for arousal and 0.436 for valence on SEMAINE.The usefulness of the proposed head and eye features has been shown in thisresearch, and they can also facilitate model interpretability efforts as the handcraftedfeatures are themselves interpretable. This work provides researchers with newaffective feature sets from video and methods that can improve affect predictionand potentially other social and affective computing efforts.
2020-07-01T00:00:00ZTowards improved trust in threat intelligence sharing using blockchain and trusted computing
https://research.thea.ie/handle/20.500.12065/4014
Towards improved trust in threat intelligence sharing using blockchain and trusted computing
Wu, Yichang; Qiao, Yuansong
Threat Intelligence Sharing is one of the most promising approaches to resist
ever-growing cyber attacks. Pressing challenges including concerns about
trust, privacy, negative publicity, policy/legal issues and expense of sharing
inhibit e ective sharing. Among all these sharing requirements, open trust issues
are most crucial and fundamental prerequisites. Blockchain Technology
has been widely adapted in data sharing market such as IoT Data sharing,
Health Information sharing and so on. Trusted Computing provides an isolated
secure trustworthy execution environment.
This thesis AIMs to investigate the use of Blockchain Technology and
Trusted Computing as a mechanism to address open trust issues of threat
intelligence sharing.
2020-06-01T00:00:00Z