Player is loading...
MHTI
Wednesday, Aug. 24, 2022 · 12:32 p.m. · 09m 49s
The objective of the project "Mental Health in Tech Industry" is to provide the general public with a prediction model based on the data from OSMI (Open Sourcing Mental Illness) of possible mental health disorders, as well as an indication to consult professionals in the area or point out improvements in the working conditions of a particular company.
Embed
Transcriptions
Note: this content has been automatically generated.
00:00:01
okay
00:00:05
okay
00:00:07
hello everyone i'm little 'cause i'm here to talk about
00:00:10
our product into him and to have the tech industry
00:00:14
an average talk about face to face with say space we have
00:00:17
tech companies to create a more mentally health environment using artificial intelligence
00:00:22
so also you can rest and well we do i was start talking about little bit but our problem
00:00:27
but that's serious mental health has been a subject discuss it more than every
00:00:30
year after year there were parts in meant to have this art isn't meant how
00:00:34
l. features had just increase it and especially with the pandemic the
00:00:38
subject has been in their minds in our conversations more than ever
00:00:42
uh it's also especially in the tech industry where more than ever in place can choose whether wanna work
00:00:48
this is our top priority for combat for companies and
00:00:51
actually a link uh dean research than quentin won't discover that
00:00:55
it has become the number one priority for in place when looking for a new job to have an environment which is
00:01:01
of values mental health has and does that work a good work life
00:01:04
balance so they are less likely to develop some kind of mental health issue
00:01:09
oh and to understand how promise the problem is it important to
00:01:13
break some data to the table so first of all according to
00:01:17
the word half are going is a shown in our century everyone in far people will develop some kind of me to have disorders
00:01:24
the two thirds of this people won't buy the won't be the animals will be
00:01:28
treated properly of beige is even bigger in the tech industry with
00:01:33
they open source mental in this organisation stating that fifty one percent of art the
00:01:38
tech industry professionals we'll be doing those with some kind of me to have desired
00:01:42
so to put it in perspective if you think about your chain during the i see if it you has far people
00:01:48
it's likely that two of them if not yourself will be that knows it was some kind of me to have deserted
00:01:55
the company is also being for companies with the word have organisations saying that depression and and say how
00:02:01
long cause global economy approximately one three million dollars we're you're due to loss protein or the c. d.
00:02:08
for a company of the size of a d. at all uh it's estimated that this is more than one point five million dollars a year
00:02:15
soul to solve this problem our solution helps companies just as hell safe and healthy they works bases
00:02:22
by providing information to identify if they imply so common
00:02:25
symptoms associated with people who have been titles with some kinds
00:02:29
for me to have disorder and then we also provide data on how the employees perceive the company actions
00:02:36
for an play side they receive a survey were quite some data
00:02:39
like demographic information personal history recent feeling is that we have in
00:02:44
and of course how they perceive the company actions and then with process this data
00:02:49
with our model of how our schiller model that has
00:02:54
been traded with thousands of data of a of a service
00:02:58
and then actually uh nowadays has an it's eight seven percent accuracy and then we give some feedback to the user
00:03:05
in this stage we highlighted the user and it's important to highlight you too bad this result it isn't meant
00:03:10
to work as a self signals is any doesn't replace the one given by a professional yeah it in any way
00:03:16
we actually give ways of offer users should connects to make the health professionals
00:03:20
so they can of delta then if they need to if they find any too
00:03:24
and our main goal here is to raise awareness on the team and break this demon
00:03:30
so with all this information with all this process data presented dashboard to the company the send the survey
00:03:36
yeah with relevance data about how the general status of the people in organisation and how they perceive the company actions
00:03:43
of this way the company can understand what's the status of the place journey
00:03:47
and creates a parking problems are in usage if she
00:03:50
will improve it's again it's importance jewel of highlights your
00:03:55
that we don't include any of the demographic questions in these these uh oh it's only used to train our model
00:04:01
and they information is only available after a sent a certain number of answers have been reach
00:04:07
so we guarantee that it's not possible to identify the individual data of each employee in anyway
00:04:14
uh the underline match here is that our solution use machine learning should which production so it's uh
00:04:19
like most current services in the market it's always
00:04:22
learning and improving itself to predicts cases with more accuracy
00:04:26
mar over our solutions not basin in single questionnaire but in a set of different opus surveys conducted by different
00:04:32
systems of research that makes it possible to have a more diverse analysis and i'll show you how the social works
00:04:41
i have the video sure
00:04:43
so
00:04:52
right so this is the interface that imply you receive this with
00:04:56
the when it is the survey there some information about harper that works
00:05:01
what the service for how the data will be collected and start and also some definitions like
00:05:06
when we use metal have bizarre then when we use mental health issue and the main difference is
00:05:11
that that the presence of a farmer diagonal this ritual of constituted is are there
00:05:16
then we have some questions that he can he or she they can answer about
00:05:21
well all things that demographic data how they're feeling recently wise is the prisoners story and also
00:05:27
what's the company bit the people work on due to them and so we perceive they are
00:05:33
of that so we can catch their perceptions of the actions the company is doing and
00:05:39
you can see we we really made a faster but the way estimate that's
00:05:43
like four minutes show answer the question we have this question with that diagnosis
00:05:47
oh also we can train our model and then there is the receivers this kind of feedback uh again we're highlighting that's not
00:05:54
up a self diagnose is that if they wished they need they feel
00:05:59
the need to talk to a professional there are other ways to do that
00:06:02
and with someone so want some information how we use these answers
00:06:07
to the company's side well we have are landing page reckon they find some information about how we
00:06:12
work while we do how we use the data and they can learn how the much the model work
00:06:19
all which kind of feedback prove get back to then be hide all the important information we already told you
00:06:25
and they can contact us if they have any doubts or any question or anything that they need to know
00:06:30
if they find it'll be our solution makes sense for the chain making that started by given us
00:06:35
some information like what organisation name uh what's the name of the employees that are currently have and
00:06:42
if you are primarily attack industry and then they create a registration for them with an email and password
00:06:48
and what they do the the registration the logging they
00:06:51
receive a link jewel sent to the employees with the survey
00:06:56
and they sent this link and they don't have access to a dashboard
00:07:00
of uh at the site ever dashboard as you can see here they can see any data has been collected because we have
00:07:06
few respondents we need more so can guarantee it's an ultimate
00:07:10
it's anonymous and once they have more more answers they can see
00:07:14
how's the distribution of respondents by the likelihood that they will be back in those of it some kind of
00:07:19
me to have to start our bayes unregulated in our tests and also how the people of perceive the company actions
00:07:27
uh each with each question there's a little explanation of how to use this data how
00:07:31
can they analyse it and what they can do to improve of their workspace based on there
00:07:38
so this is our them all how do i turn back
00:07:44
yeah
00:07:46
i don't know all right uh this thirteen has been working show brings a special lice
00:07:53
camilla is r. s. q. for you x. y. designer their guarantees we have i
00:07:57
didn't really experience in our product and that we treat the subject in a respectful way
00:08:02
the new louisa who's text after engineer who can bring save space
00:08:06
why should and create half full of uh experience for our users
00:08:12
and i am an h. i. professional has been studying well being in the workspace
00:08:16
for the last year is and then we decided to to big attack approach to it
00:08:21
oh talking about our next that we wanna keep training the model our model
00:08:24
is going to financial with all the questions that we collect except for the ones
00:08:29
about recent feelings because we each have a larger data set for that so we'll
00:08:33
plan on train there have that uh as we collect my answers for non users
00:08:37
we also wanna build partnerships with companies that make the connection because we between people and health professionals
00:08:44
oh we also be a wanna be able to get companies mar recommendations
00:08:49
about what they can do based on the data we receive
00:08:51
from then a similar combinations we simply companies with this in context
00:08:57
and one incorporate mar service in our model so that like the flowers you mean the aches and
00:09:03
came at organisational climate service so that we can have a mark personalise herbal products should which companies needs
00:09:10
we've been doing that and working that in the next most believe six mark moans
00:09:15
i like to people they're already talking about one with uh with with a devil upper back around them with the
00:09:21
business background we can keep of beauty in our product and bring all this new uh stuff that we just showed you
00:09:30
oh uh we are very happy and very uh we're proud of our which even this hack it
Conference Program
LightAI4Comfort
Aug. 24, 2022 · 12:13 p.m.
119 views
LaternaMagica
Aug. 24, 2022 · 1:08 p.m.
118 views
SolveHR: Natural language inference for professional search
Aug. 24, 2022 · 1:26 p.m.
Recommended talks
L’analyse de données et ses bénéfices pour le patient
Immo Oliver Paul, fondateur de Carenet (Brésil)
June 6, 2014 · 1:03 p.m.
Embedded Interactive Learning Dashboards with ElasticSearch And Kibana
Andrii Vozniuk, EPFL
Sept. 21, 2016 · 8:17 p.m.
481 views