Sangeetha Mohan
10/13 itu question fottnoted
great tal adam fysh - hear are biggest unanswred questions in risk - how can ai help
This webinar will discuss the potential role of AI in reducing disaster risks in vulnerable countries and globally, and thus advance implementation of the Sendai Framework for Disaster Risk Reduction (DRR) and Sustainable Development Goals. It will examine how AI applications can support more efficient and effective decision-making across all stages of disaster management, from early warning and monitoring to disaster preparedness, emergency response and post-disaster recovery/relief.
There will be focus on how AI can help transform thinking and behaviour towards disaster prevention and contribute towards risk-informed, sustainable development and investments.
note which un subnetworks hold which master risk framework
constraints data sharing or model algorithm
Thats a nice initiative to have AI usage in the risk reduction and management process related to natural disasters. But do we have tight data security , data & AI ethics & compliance rules in place for its usage ?? Without stringent AI ethics in place, the adverse impact would be of much higher magnitude when this AI model is REACHED AT THE WROND HANDS, and leveraged to target specific areas of the planet and mimic & simulate the natural disasters and related risks. Its very important to ensure the AI ethics &data security and governance measures in place as we are taking a huge dataset,the whole nature itself for this research study and implementation.
HUAWEI P20 lite 05:35 AM
Will this data be free to access by everyone or Google is going to build just another business around them?
Sella Nevo 05:53 AM
Great question. This information is provided for free and made publicly available. You can see current live alerts in g.co/floodalerts. But we're not only making our outputs publicly available but also the models themselves. For example, you can see some of our open-source models in this paper: arxiv.org/abs/1907.08456. We're dedicated to making life-saving technologies available to everyone in the world, for free.
jomon 05:14 AM
How data collection for disaster prediction can be done without affecting privacy ?
Bilel Jamoussi 05:42 AM
Significant amounts of data are NOT Personally identifiable information (PII) - e.g., satellite images, telecom connectivty maps, hydrolic waterways, etc. This non PII data is critical in AI based DRR
Sangeetha Mohan 06:11 AM
Here PII data would n't be an issue per say wrt the data security. But may be we would have to coin a new word say, location identifiable information for protecting the privacy and critical identifying details of a particular geography/ location.
Gautham Bejoy 05:18 AM
will it be possible to make a robot which can use the idea of geolocation in rescuing people who get trapped beneath the soil during landslides?
Dr Keshav Mohan 05:23 AM
How can AI capabilities and Big Data insights be applied to ensure educational rights to children particulary in prolonging disaster scenarios? Is there any best examples for it currently in the circumstances that education facilities are badly afftected most often during any disasters and that SFDRR identified it as a core area for actions.
Dr Keshav Mohan 05:33 AM
Sella Nevo: Is there any oppotunity for Enggineering students to partner with DRR inititives of Google?
Sella Nevo 06:00 AM
Of course! I think the easiest way to collaborate closely is through internship positions. We've had multiple internships at Google Research where students were able to not only work at Google in the summer on projects unrelated to their academic work, but also work on academic projects that are of interest to both Google's humanitarian work and their adviser's interest, which I believe is a big win-win.
Dr Keshav Mohan 06:14 AM
Thank you so much. Can you share links to contact by the students / faculty for the purpose.?
Sunil Prabhakar 05:34 AM
Sella, do you have any studies on 2018 Kerala flood?
Saiful Ridwan 05:36 AM
Question to @sella: Is the ML model applicable for cities where there many man made structures, both increasing and decreasing the flood impacts ? Are these dimensions factored in the model ?
Sella Nevo 05:55 AM
This is an important point. As I tried to really briefly mention during the slides (sorry I didn't have time to get into it more), man-made structures are one of the more challenging factors to address at scale. We've partially begun to deal with these, taking into account dams and embankments, but there are still many man-made structures (particularly in cities like you've highlighted) which we still don't model. In areas where flooding patterns are significantly affected by such structures we do not (yet) launch our systems, rather than launching models which erroneously ignore them.
Saiful Ridwan 05:59 AM
Thanks for the response. Appreciated.
Webinar Attendee 05:36 AM
I missed the first part of the flood modelling presentation. What is the lead time?
Chat and Q&A 05:40 AM
Hello, you may find a recording of the session after today on https://aiforgood.itu.int/events/how-can-artificial-intelligence-and-digital-technologies-reduce-disaster-risks-in-countries/ or https://youtu.be/jmYmZLNatro
Webinar Attendee 05:54 AM
Thanks a lot.
Anika Bijil 05:38 AM
Well... I reside in the Middle East. And often flood alerts turn out to be false. So... Can we possibly say that the AI used in not accurate enough?
Sreebala Ajith 05:39 AM
Shall I get information regarding the Online/Offline courses that could be done in the field of Artificial Intelligence?
Dr Keshav Mohan 05:59 AM
Coursera gives a number of courses. SNIT ADOOR has a campus agreement with coursera that you can do courses free and get certificate.
Sreebala Ajith 06:03 AM
Thank you sir for the reply.
Vandana Yadav 05:40 AM
Will the recording for this session be available later?
Chat and Q&A 05:41 AM
Hello, you may find a recording of the session after today on https://aiforgood.itu.int/events/how-can-artificial-intelligence-and-digital-technologies-reduce-disaster-risks-in-countries/ or https://youtu.be/jmYmZLNatro
Malek Jebaie 05:41 AM
AMAZING Presnetation Sella !how can google data be used? Do we need to have sharing data agreement? or it is an open data ? and it is possible to have access to the power point material ypu presented?
Sella Nevo 05:58 AM
Hi Malek, thank you! All of our alerts are publicly available at g.co/floodalerts. If you need to automatically ingest our alerts, you can look for details on on ingesting Google Public Alerts through our programmatic APIs, or reach out to me and I'll connect you to the relevant people.
Malek Jebaie 06:02 AM
Can i have your email please?
Webinar Attendee 05:42 AM
are you monitoring the floods and rainfalls in Kerala?
Sella Nevo 06:05 AM
In Kerala, we are not yet involved in producing the forecast itself (which is created by the Indian Central Water Commission), but we use our inundation models to identify who needs to be alerted, and use our distribution systems to make sure more people receive this information in a timely matter. We have only started working in Kerala a few months ago, so there's still a lot we want to do there which we're not yet doing.
Webinar Attendee 06:09 AM
good to hear you are here, it'll be much helpful for the people if these technologies are with them.
Subhash Chandran 05:45 AM
How challenging it would be for the AI research to predict the disasters in newly developed/developing cities/towns especialy at environmentally fragile zones ? As we have seen in the the recent years that the severity of the natural disasters has been aggravated by the poor city/town planning resulting in the loss of life and property. Hence, what should be the ideal strategy to integrate AI to the early stages of city planning?
Bilel Jamoussi 05:52 AM
Please see the ITU publication on Disruptive technologies and their use in disaster risk reduction and management https://www.itu.int/en/ITU-D/Emergency-Telecommunications/Pages/Events/2019/GET-2019/Disruptive-technologies-and-their-use-in-disaster-risk-reduction-and-management.aspx
Sunil Prabhakar 05:46 AM
@ Preeti Banzal: AI used for Kerala and Chennai floods? Would you be able to provide more details on it?
This question has been answered live
Sunil Prabhakar 06:14 AM
Requested for further details too
Vyshnavi Venu 05:59 AM
WHAT WILL BE THE MAJOR CHANGES NOT ONLY IN FIRLD OF DIASTER RELEIF BUT ALSO IN THE FEILD OF EDUCATION AND HEALTH
Rudolf Dema 06:02 AM
Thank you Yves for the fantastic presentation. Could you please provide the name of the african forum on the drone mapping
This question has been answered live
Yves Barthelemy 06:09 AM
https://www.africandroneforum.org/
CARLO VERMIGLIO
Does speakers have presentations/slides available in order to better follow the discussion?
Joanna Batstone 05:52 AM
I would be happy to provide my speech notes and references if that would be helpful too
Smitha Kallukadavil Mohandas 05:54 AM
I need the links
CARLO VERMIGLIO 05:57 AM
Yes please, it would be great.
CARLO VERMIGLIO
How can AI and digital technologies contribute to the performance management of a local government in disaster and emergency scenario?
Webinar Attendee
Thanks for the presentation very on-point thus far. Excellent stuff.
Bilel Jamoussi 05:37 AM
You are welcome. Thanks for joining!
Ajo Aloysius 05:12 AM
Can you explain the assessment that is used to test the intelligence of a machine.
Bilel Jamoussi 05:45 AM
Typically ML models are ran against test data sets to evaluate their accuracy; domain experts do the evaluation and feedback
Francesco Alesiani 05:16 AM
How will Monash University provide access to dataset for research and validation?
Joanna Batstone 05:49 AM
The Victorian vegetation map I referenced can be found here - http://francois-petitjean.com/Research/MonashVegMap/map.html
Francesco Alesiani 05:51 AM
thank you
Ashraf Bava 05:22 AM
How people can benefit from the Toll edveloped by Monash university for Kerala?
Preeti Banzal 06:01 AM
yes they are very well documented and available as public information. thx
Chat and Q&A 05:26 AM
From Chitra Kumar: "Do Monash university coordinate with kerala Goverment about the findings of flood model prepared for the use of man kind"
Joanna Batstone 05:40 AM
A nice summary of our work in Kerala can be found here - https://lens.monash.edu/@technology/2020/07/07/1380771/satellite-saviour-new-mapping-technology-to-help-india-prepare-for-floods
Gagandeep Singh 05:39 AM
Question directed to Sella Nevo: Have you worked on flash flood forecasting especially in the Northwestern Himalayan Region. Can you share some insights on the same. The India map shown in the 2nd last slide showed rivers of Uttarakhand (Ganges). What data sets is google using for flood warning?
Sella Nevo 06:03 AM
Hi Gagandeep, thanks for the question. Our work currently focuses on riverine flooding, and we do not support flash flood warnings yet (we hope to work on this in the future).
In our systems we use a wide range of in-situ measurements, satellite-based measurements, and weather models. In many of these we rely on the data and support of the governments we work with, including the Central Water Commission in India and the Bangladesh Water Development Board.
Md Afjal Hossain 05:41 AM
May we receive the recording of this webinar?
Chat and Q&A 05:41 AM
Hello, you may find a recording of the session after today on https://aiforgood.itu.int/events/how-can-artificial-intelligence-and-digital-technologies-reduce-disaster-risks-in-countries/ or https://youtu.be/jmYmZLNatro
Dr. Arshad Khalafzai 05:41 AM
Vulnerability: Physical exposure to the hazard risk, sensitivity of the impacts, and adaptive capacity (both perceived and material), inlcuding coping capacity, or resilience of individuals. So, adaptive capacity indicates long-term capacity or resilience as opposed to coping capacity, which is short-term capacity to deal with a hazardous event.
Krishna Neupane 05:42 AM
Association of Disgtrict Coordination Committees of Nepal would like to invite Monach University for join action to research the Mountain climate and strategy for the mitigation of risk? please communicate : kneupane47@gmail.com or adccn77@gmail.com
Gaurav Nemade 05:42 AM
How can AI and ICT help in the reduction in the spread of epidemics (health disaster) and risk mapping in order to minimize its socio-economic implications?
Malek Jebaie 05:43 AM
How can development countries benefit from AI applications if such countries have issues in data generating and centralization?
Dr Keshav Mohan 05:43 AM
Preethi Banzal ji: Is there any best examples to cite whereby AI is applied ensuring lastmile connectivity & early warning in India
Hareesh N A 05:48 AM
Sella Nevo: The flood warning of Google indicates the risk areas of innundation and rising level of rivers nearby. But the perception of risk to a common man also needs to be broought in.What does a common man perceive the flood risk to be ? what actions needs to be done if the area is in danger? so will these be integrated in future ?
Sella Nevo 06:10 AM
This is an excellent point. Translating the facts on the ground into risk, and conveying this risk to the general population is a major area of research for us, which we're doing in collaboration with Yale university and the World Bank. The information we produce is already addressing requests we receive from affected individuals on the ground based on multiple field surveys we've done, but there's still a lot of room for improvement.
Regarding actions that need to be taken - this is also a critically important point. We already provide recommended actions and safety tips below the forecast itself (not sure if it could be seen in the slides), but any action we recommend taking has to be approved and provided to us by the relevant government agency - so often updating these and making these tailored to the current situation can be challenging.
Francesco Alesiani 05:48 AM
Would be possible have the references of the work from Sella Nevo?
Malek Jebaie 05:50 AM
yes please
Sella Nevo 06:12 AM
Gladly!
Here are the references that appeared in the slide deck:
arxiv.org/abs/1907.08456
arxiv.org/abs/2007.00595
https://ai.googleblog.com/2020/09/the-technology-behind-our-recent.html
We've published a lot more work that wasn't referenced in the slides, feel free to search for Google Flood Forecasting or ask about work you're particularly interested in here.
Francesco Alesiani 05:50 AM
@ Preeti Banzal: how do foreseen interation of people with AI, in relation to the different languages?
UNNIKRISHNAN PS 05:53 AM
Sir, what are the current technologies to predict natural disaster?
Massoud Hamad 05:53 AM
@Yves, Can you talk this topic in the Context of Zanzibar
Krishna Neupane 05:59 AM
Dear Yues, thank you for the wider range of information about Disaster and support of WB. I am urging WB through you to support Nepal on planning to the poor and vulnerable poeple living in unsafe high land which was wratched by divastating earth quake in 2015 and it is further risking by heaving rain after landslide. You can explore for the special research on this and my association which is umbralla of all 77 districts coordination Committees representating by elected local leaders. what is the possibility to work in this area using AI for good. It would be democratic approach for DRR in sub national Government level. www.addcn.org.np
Jasmine Joseph 05:59 AM
Heard that china is using AI for tracking and reporting Covid19 cases. How can a similar strategy be applied to economically challenged communities. Is there a way to use IoT devices which are cheap enough ..
Faith Amatika-Omondi 05:59 AM
thank you so much for your presentation. could you kindly share more information about the project in Zanzibar and other parts where machine learning and AI has been involved?
Yves Barthelemy 06:06 AM
https://www.opendri.org/resource/machine-learning-for-disaster-risk-management/
Anilkumar S 06:01 AM
India has invested substantially in dams, how can AI be associated with dam management and safety?
Rajendra Kumar Agrawal 06:01 AM
in this time of uncertainty and climate change how data can be standardized??
Joslin Jose 06:01 AM
how can AI be used to assess climate change ?
Jum'atil Fajar 06:01 AM
To Joanna - Do you have publications about your work in Indonesia?
Joanna Batstone 06:03 AM
Monash has a number of projects underway in Indonesia - some examples are referenced in these two links - https://lens.monash.edu/@environment/2019/03/22/1374017/bringing-new-life-to-settlements-in-makassar
and
https://lens.monash.edu/@environment/2020/06/25/1380719/lets-radically-rethink-the-relationship-between-our-environment-and-human-health
UNNIKRISHNAN PS 06:01 AM
Is Artificial Intelligence likely to become a subject of study in the future?
Anika Bijil 06:02 AM
Sella: The more steep in the graph the more better. If the land itself is quite steep, won’t it affect the graphs?
Sella Nevo 06:14 AM
Hi Anika, I'm not sure which graph you're referring to here. Are you referring to the graph comparing performance of the LSTM-based models to classic hydrologic models? Could you elaborate on the question?
Jaison. Kappil 06:05 AM
my question is that, After the Corona period, do you plan to formulate any plan for the people of the world?
Krishnakumar k v 06:08 AM
villages experiencing netconnectivity problems, how it can be solved?
Daniel Haltmeier 06:09 AM
Several panelists have talked about satelite imagery analysis with the help of AI. However, this is a classical dual use problem and the exact same software could be used for military purposes. Is there a discussion of this point in the scientific community developing these technologies for disaster relief and if yes, could someone explain?
Egide Nijimbere 06:09 AM
Although AI is growing up along with good and remarkable innovation, developing countries are still scared to incorporate them in their systems due to security threats. Is there a way of assuring them as natural disasters had became a world concern from which everyone have to contribute? Thanks
Dr Keshav Mohan 06:13 AM
Thank you so much. Can you share links to contact by the students / faculty for the purpose.?
Krishna Neupane 05:59 AM
Dear Yues, thank you for the wider range of information about Disaster and support of WB. I am urging WB through you to support Nepal on planning to the poor and vulnerable poeple living in unsafe high land which was wratched by divastating earth quake in 2015 and it is further risking by heaving rain after landslide. You can explore for the special research on this and my association which is umbralla of all 77 districts coordination Committees representating by elected local leaders. what is the possibility to work in this area using AI for good. It would be democratic approach for DRR in sub national Government level. www.addcn.org.np
Jasmine Joseph 05:59 AM
Heard that china is using AI for tracking and reporting Covid19 cases. How can a similar strategy be applied to economically challenged communities. Is there a way to use IoT devices which are cheap enough ..
Faith Amatika-Omondi 05:59 AM
thank you so much for your presentation. could you kindly share more information about the project in Zanzibar and other parts where machine learning and AI has been involved?
Yves Barthelemy 06:06 AM
https://www.opendri.org/resource/machine-learning-for-disaster-risk-management/
Anilkumar S 06:01 AM
India has invested substantially in dams, how can AI be associated with dam management and safety?
Rajendra Kumar Agrawal 06:01 AM
in this time of uncertainty and climate change how data can be standardized??
Joslin Jose 06:01 AM
how can AI be used to assess climate change ?
Jum'atil Fajar 06:01 AM
To Joanna - Do you have publications about your work in Indonesia?
Joanna Batstone 06:03 AM
Monash has a number of projects underway in Indonesia - some examples are referenced in these two links - https://lens.monash.edu/@environment/2019/03/22/1374017/bringing-new-life-to-settlements-in-makassar
and
https://lens.monash.edu/@environment/2020/06/25/1380719/lets-radically-rethink-the-relationship-between-our-environment-and-human-health
UNNIKRISHNAN PS 06:01 AM
Is Artificial Intelligence likely to become a subject of study in the future?
Anika Bijil 06:02 AM
Sella: The more steep in the graph the more better. If the land itself is quite steep, won’t it affect the graphs?
Sella Nevo 06:14 AM
Hi Anika, I'm not sure which graph you're referring to here. Are you referring to the graph comparing performance of the LSTM-based models to classic hydrologic models? Could you elaborate on the question?
Anika Bijil 06:21 AM
Oh yes, the LSTM Graph.
Anika Bijil 06:22 AM
In a quite sleepy hill, do you think the steepness would also affect the LSTM Graph?
Jaison. Kappil 06:05 AM
my question is that, After the Corona period, do you plan to formulate any plan for the people of the world?
Krishnakumar k v 06:08 AM
villages experiencing netconnectivity problems, how it can be solved?
Daniel Haltmeier 06:09 AM
Several panelists have talked about satelite imagery analysis with the help of AI. However, this is a classical dual use problem and the exact same software could be used for military purposes. Is there a discussion of this point in the scientific community developing these technologies for disaster relief and if yes, could someone explain?
Egide Nijimbere 06:09 AM
Although AI is growing up along with good and remarkable innovation, developing countries are still scared to incorporate them in their systems due to security threats. Is there a way of assuring them as natural disasters had became a world concern from which everyone have to contribute? Thanks
This question has been answered live
You 06:14 AM
what is role of "global" professions in opening up top secret risk data to ai? i attended many professional risk summits debating the then big 5 accountants around 2000- every profession was hiding risk it knew most about- eg insurers dont publish what caused meltdown we need to minimise current liabilities not prevent reoccurence
Marc Liberati 06:15 AM
As to Adam's final point on capital flows, is the next step in DRR embedding consensus risk models directly into monetary policy and allocation of investment capital for a green and risk reduced future in furtherance of the SDGs.
Adam Rowland Fysh 06:22 AM
It can't be forced, but we can certainly help them. We've heard plenty of development finance organizations, insurance companies and donors interested in global public good and taking portfolio approaches to investing in risk reduction and we'd like to help them with principles, tools and capacity building to make it easy for them.
Francesco Alesiani 06:16 AM
@Fysh: how do you think AI could help in the Risk Reduction process?
MANISH RANJAN 06:16 AM
ANY COMMENTS ON THE ETHICAL ISSUE OF AI IN risk assessment
Massoud Hamad 06:18 AM
@Yves, I need to talk about this topic in the context of Zanzibar, please can you give me your mobile contact
Sreya Vaidyan 06:20 AM
How can we empower urban local bodies and provide capacity building to the local government when we introduce AI?
Yves Barthelemy 06:25 AM
maybe you can have a look at some of the blogs here: https://opendri.org/category/news/ they could give you ideas on how to get started ?
Akhil Jose 06:21 AM
Will the ML algorithm and API be available in public domain?
Egide Nijimbere 06:22 AM
Agreed that collaboration, sharing experience and partnership is the most, but have think how this is going to be implemented on the community lever for their resilience? Thanks
Yves Barthelemy 06:26 AM
Have a look at these blogs, they will show you some examples on how to engage with communities on this specific resilience issue https://opendri.org/category/news/
Cyrus Kiambati 06:24 AM
I am not an expert in the hydrology field... I however happened to carry out an environmental assessment in for a road project in the Northern part of Kenya. From my observations, the project is deemed to fail as the hydrological pattern and characteristics of the area was never considered in the designs. The effects of this is currently being experienced by the host community. What are your recommendations?
Webinar Attendee 06:28 AM
I have a query sir. Can we develop a system which connects our police stations ( I am emphasising police stations because they are the most popular and common government offices which relates to common people.) and these disaster warning systems. Because It will be more effective especially in local areas for example high range areas. May be it help to save more people( here I am thinking about the 'Pettimudi' incident in kerala recently, where we were not able to save majority).
No comments:
Post a Comment