0
Hopp til hovedinnhold

Bruk av kunstig intelligens for ledelsen ved brann-katastrofer

Jivitesh Sharma fra Fakultet for teknologi og realfag disputerer for ph.d.-graden med avhandlingen «Advances in Deep Learning Towards Fire Emergency Application: Novel Architectures, Techniques and Applications of Neural Networks» mandag 11. januar 2021. (Foto: Privat)

The main objective of this thesis is to advance the state-of-the-art methods in Artificial Intelligence and apply them to the important task of Emergency Management. Specifically, we focus on fire emergencies as they are the most common type of hazard.

Jivitesh Sharma

Ph.d.-kandidat

Disputasen foregår digitalt på grunn av korona covid-19-situasjonen. Se nederst på siden for hvordan publikum kan overvære disputasen.

 

Jivitesh Sharma fra Fakultet for teknologi og realfag disputerer for ph.d.-graden med avhandlingen «Advances in Deep Learning Towards Fire Emergency Application: Novel Architectures, Techniques and Applications of Neural Networks» mandag 11. januar 2021.

Han har fulgt doktorgradsprogrammet ved Fakultet for teknologi og realfag, med spesialisering i IKT.

Slik oppsummerer Jivitesh Sharma selv avhandlingen:

Advances in Deep Learning Towards Fire Emergency Application: Novel Architectures, Techniques and Applications of Neural Networks

The main objective of this thesis is to advance the state-of-the-art methods in Artificial Intelligence and apply them to the important task of Emergency Management. Specifically, we focus on fire emergencies as they are the most common type of hazard.

The problem is divided into three stages: Detection, Analysis and Evacuation Planning. We provide Artificial Intelligence based solutions for each stage in a disjoint manner.

Detection:

Current systems lack the high precision of Artificial Intelligence based methods and require more hardware such as smoke detectors, thermal detectors etc.

For detecting an emergency, we use two ways: visual and audio recognition. Novel Artificial Intelligence based computer vision techniques are proposed to detect fire with high accuracy, that might potentially be used with camera systems that we have today.

For audio-based detection, we propose a novel sound recognition model using neural networks that sets the new state-of-the-art in this application. The auditory emergency recognition model is a general-purpose emergency detection system.

Analysis:

We propose the first Artificial Intelligence based emergency analysis tool to thoroughly analyse a disaster-stricken area. A neural network based object detection and segmentation model is proposed. The model distinguishes and segments objects in the emergency environment based on their build material to convey information about the objects’ vulnerability to catch fire. Also, people are also detected and segmented to get a rough head count and location. The model could also provide a rough direction of fire spread.

This can be extremely crucial information for the search and rescue personnel that usually go in without any useful information, to save people. This can potentially reduce the risk of injury or death for the fire fighters and help in rescuing people more efficiently. 

Evacuation:

For the final and probably the most important stage of the emergency management procedure, we propose the first full scale evacuation planning model based on Artificial Intelligence. Specifically, we use Reinforcement Learning to plan optimal evacuation strategies to evacuate all people inside a building. We include various realistic features and scenarios in our simulator including dynamic fire spread, bottleneck (max. number of people inside a room/hallway), uncertainty (to model human behaviour), timeliness constraint (to evacuate everyone in the least amount of time), multiple fires etc.

We train a Reinforcement Learning agent on this simulator with the objective of evacuating everyone in the least amount of time while avoiding any hazardous areas in the building. In order to scale our model to work on large buildings, we employ attention based Reinforcement Learning. We show the effectiveness of our method by running a simulation for fire evacuation on our own UiA building in Grimstad. Our method is able to evacuate nearly 1000 people from the UiA building in nearly optimal time, without putting anyone in harm’s way. We also provide mathematical guarantees for our model

 

Disputasfakta

Prøveforelesning og disputas finner sted digitalt i konferanseprogrammet Zoom (lenke under).

Disputasen blir ledet av førstelektor Morgan Konnestad, Institutt for informasjons- og kommunikasjonsteknologi, Fakultet for teknologi og realfag, Universitetet i Agder.

Prøveforelesning: 14:15
Disputas: 16:00

Oppgitt emne for prøveforelesning«Ensembles and the bias / variance dilemma»

Tittel på avhandling: “Advances in Deep Learning Towards Fire Emergency Application: Novel Architectures, Techniques and Applications of Neural Networks

Søk etter avhandlingen i AURA - Agder University Research Archive, som er et digitalt arkiv for vitenskapelige artikler, avhandlinger og masteroppgaver fra ansatte og studenter ved Universitetet i Agder. AURA blir jevnlig oppdatert.

Avhandlingen er tilgjengelig her:

https://uia.brage.unit.no/uia-xmlui/handle/11250/2721827

eller her:

KandidatenJivitesh Sharma (1991, New Delhi, India) Bachelor of Technology fra Guru Gobind Singh Indraprastha University, New Delhi, India (2013), Masters degree in Technology fra Maulana Azad National Institute of Technology, Bhopal, India (2016).

Opponenter:

Førsteopponent: Førsteamanuensis Paul W. Munro, School of Computing and Information, University of Pittsburgh, USA

Annenopponent: Professor Heri Ramampiaro, Institutt for datateknologi og informatikk, NTNU

Bedømmelseskomitéen er ledet av professor Frank Reichert, Institutt for informasjons- og kommunikasjonsteknologi, Universitetet i Agder

Veiledere i doktorgradsarbeidet var professor Ole-Christoffer Granmo, UiA (hovedveileder) og professor Morten Goodwin, UiA (medveileder)

Slik gjør du som publikum:

Disputasen er åpen for alle, men for å følge prøveforelesning og disputas må du melde deg som publikummer på denne lenken:

https://uiano.zoom.us/meeting/register/u5UvcOmqrzIrGdL04nVc2EKJ_1KiDFhn_SCO

Du får en Zoom-lenke i retur.

(Her er framgangsmåten for å bruke Zoom: support.zoom.us om du ikke kommer inn ved å klikke på lenken.)

Vi ber publikum om å ankomme digitalt tidligst ti minutter før oppgitt tid - det vil si til prøveforelesningen 14:05 og disputasen tidligst 15:50. Etter disse klokkeslettene kan du når som helst forlate og komme inn igjen i disputasen. Videre ber vi om at publikum slår av mikrofon og kamera, og har dette avslått under hele arrangementet. Det gjør du nederst til venstre i bildet når du er i Zoom. Vi anbefaler å velge «Speaker view». Dette velger du oppe til høyre i bildet når du er i Zoom.

Opponent ex auditorio:

Disputasleder inviterer til spørsmål ex auditorio i innledningen i disputasen, med tidsfrister. Disputasleders e-post er tilgjengelig i chat-funksjonen under disputasen. Spørsmål om ex auditorio kan sendes til disputasleder Morgan Konnestad på e-post morgan.konnestad@uia.no