Intro to ETSI 3GPP Mission Critical Services (MCS)

You might not know it, but there's a whole world of service standards like Professional Mobile Radio/TETRA and similar that exist and are being used in the world around us that general IT community is not closely aware of.
We've had a chance to do two ETSI Plugtests lately and are developing a MCS product landscape around 3GPP standards and are involved in the standardization procedures.

Structure of the session would be:
- share 3GPP views and some words on ETSI
- present what MCS actually means with some real world examples (e.g. in police/fire departments, military, retail, transport/railways, mobility, there's a multitude of examples).
- check what 3GPP has to say on this - various releases, and levels of functionality offered
- what's the architecture of an MCS system (mention and explain all building blocks)
- what are the functions and how MCS works
- share experiences on plugtest / tell what we're allowed to say (format, setup, pretest session, testing, tools, only publicly available information will be shared because of NDA and ROE)
- GIve an outlok how MCS Future Railway Mobile Communication System (FRMCS)
- show a live demo of our system on several mobile devices (incl. some special ones)

What will be also touched during the session is
- protocol level information (e.g. protocols are used - sip/rtp)
- audio/video encoders (AMR-WB, H.264)
- realization of such networks/systems (virtualization, containerization, usage over LTE)

Changing the world (of traffic control)

Why do we IT geeks and geeketes do what we do? It could be as simple as really just delivering what the customers want of us and rarely do you get a chance to do "something that matters". So, this session would be:
- focusing around the technology of how we accomplished things with Azure,
- giving details and evolution from where we were to where we're going (and having MS cloud architects on board),
- fuzzying a bit about ML and what's the complexity here, no solutions due to NDA :)
- for the duration of the talk we can drive a simulation of traffic in some part of Osijek on the wall via projector (for example Strossmayer street as a clear major street with some side-streets, e.g. the ones that go from Lidl on the west side down to the venue and off to the center of Retfala), this will give an interesting background that will also be a point to discuss (what's up with simulation for ML and learning modelling)
- I could do chalk and talk approach and sketch the architecture evolution and give hints what was good, what could be improved, where were the major issues and what we were thrilled with - I won't go into too deep technical details of the solution but since we did it in DevOps (former VSTS) with automated deployments, this could also be a side-part of the story (DevOps approach in general)

Building AI systems that can change the world

Talk is about building new UX over AI modules targeting seamless, non-intrusive, touchless and accessiblity oriented system with various twists, hackatons, awards and variants of app that happened as a consequence of current pandemic.
One could say - instead of a very "boring" year, this turned out to be one really filled with events and accomplishments - and it's still going.

Presentation is conceptually built in stages, more like layers of opening up the story and adding sugar on top as we go till the end.
1) Intro
- where presenter(s) come from
(set the stage, a short "mandatory" mention of Atos, firm where this happens with a few sentences)
- how this all came to be?
- work with students, mentoring focus, building AI know-how, long term work on AI projects
- do some basic intro into AI
- explain what an AI project looks like (and could/should)
- how much of AI project is actually AI related
2) Talk about UX / UI
- focus on output side
=> explain human cognitive abilities (e.g. expected numbers of neurons involved into understanding a picture),
=> visual and audio focus, difference between static and dynamic (video) pictures on screen
- focus on input side
=> how we do things typically (keyboard, mouse, touch screen, desktop/laptop/handtop-erm-held)
=> changes when it comes to audio and video input (aka microphone and camera with or without temporal aspect)
=> explain text2speech, speech2text - general audio frameworks one can use from start - how it typically works out
=> explain picture pattern detection (aka video frame per frame) + typical picture classification approaches
=> explain temporal video recognition shortly
=> talk about performance (aka do I need a few GPUs to get this going?)
3) What happened in the world and why we need this tech?
- where we were at March this year?
=> gesture recognition examples
=> audio recognition examples
=> our internal mood barometer app example
- what happened
=> covid-19 = no more touching!
(sharing any screen, mouse, keyboard is a big no-go!)
=> hackatons = help the world, try to put this tech to good use
=> accessibility angle worth noting
=> where we go from here
- outlook of the new normal, how this tech is going to help us in the future
- retrospective of lessons learned
4) live-demo of one of our apps on stage on a laptop (a normal dev PC, no GPU, no fancy stuff)
5) Q&A / Discussion

(User) Interfaces - What's AI got to do with it?

Behold - you'll learn that you're always tracked, metered and evaluated.
You're a user of many systems that surround you and you're using those (effectively being a citizen/user) sometimes even without noticing it or knowing about it.

Now is the time when principles of life as we know it have shifted so profoundly that we can't humanely even track it. As software vendors, we're collecting more and more data - with Computer Vision and Narrow AI becoming more effective and able to handle tens of thousands of video-streams live and in parallel we seem to be nearing a utopia or is it perhaps a dystopia?

How is all this related to (user) interfaces?

This talk will have a classical structure:
- intro and accept idea of interfaces all around us
- where are we now, what's the state of tech from the past (incl. sci-fi) all the way to how things are and what do we as big-IT companies do with all available data
- we'll provide details on how each interface old and new can be augmented or used to paint a bigger picture and how can that help humanity
- provide some real-life examples and give you an insight into how we try to use this for good
- conclusion: is all this a good or a bad thing will be discussed with motivation to use this for the good

Talk is based on years of experience on building large-scale enterprise grade systems from telecommunication domain down to individual-user applications all of which also could've or has in fact used some form of AI (narrow ML algorithms) to generate value for users or ventures running such systems.

Motivation is to enhance people's views in as much as possible light technical discussion (meaning trying to keep it high-level and with lots and lots of visual examples).

Similar talk has already been given on "Bounding Box 2022" conference in Osijek.
Talk can be given in English and Croatian, in chalk-and-talk format (if necessary without all the fancy presentation graphics) or in beamer based format aka classical conference setting.
It can also be light or heavy given the track it's assigned to (i.e. deep dive on Computer Vision, example models, frameworks, application demonstrations, backend related big-telco ML models for various purposes, AI in general with user interfaces being the fore-front of next step in digital revolution)