0:01 I want one of these in my classrooms every day, you know. That would be really cool. Can you hear? Okay. Hi. Good morning everybody. My name is David Gerber. I am a faculty member at the University of Southern California. I do want to say thank you all for coming to listen to me and also to our hosts of course. It’s nice to see some friendly faces and it’s nice to meet some new people.

0:29 I’m going to start with going back about seven years. I was you know working at Arab with the research team and we did this project on digital twins. How many people hear digital twins and want to grown? Okay. Well, I mean it’s a little bit of an inflated term. And I came back to California from London and I was working with some great people who you will meet shortly.

0:58 With the idea that we wanted to create a campus twin, like every single asset that we could get for this university, every bad BIM model that’s been produced over the last, I don’t know, 30 years of of different fidelities. And we started to put together this hypothetical project which was a campus twin with the goals really of showing the university how they could move towards net zero.

1:20 USC is a very large organization. It’s 60,000 employees as the largest private employer in Los Angeles County. It’s a big thing with multiple campuses. So we set upon basically building this u campus twin. This was a nonfunded project. It was it was partly related to what my colleagues at tracks and and my developers were working on which is their startup and what it inevitably proved out to be was a pivot away from building digital twins for commercial commercial reasons.

However, what happened to me in that that sort of period of time was the university was building a new building. This is the new building. It’s actually completed. It’s the school of advanced computing. So all of these universities are competing to be you know the AI schools of course and our computer science is actually quite strong. And as part of the brief which had nothing to do with me it had everything to do with the the trustees of the university.

2:21 They they put these words out there that the building was supposed to be a living lab. And the dean of engineering came to me and said well maybe you could figure out what that really means for us. Yeah, quite quite literally. So I went around and I had these workshops with faculty and all the faculty rolled their eyes and said, “You’re wasting our time. This is never going to happen.

2:45 You’re not going to have any money. There’s no sensors.” D and I got a few people to give me some good ideas. At the end of the day, that this whole message really is about turning a building into an instrument of science. And you know when you think about the university, our number one asset really is what? People. Thank you. It’s our people, right? But we we always ignore our real estate assets, our buildings as assets, but we don’t ignore our MRI machines, our electron microscopes, our our nuclear, you know, testing equipment, but we never treat the building as an instrument of science.

3:20 So basically, this is a pretty traditional process. A bunch of architects competed for the work. H were the architects that won the work. They went through a whole series of design process negotiating with all of the stakeholders and eventually you know they chose a specific scheme with a with a specific mandate though as well as it had to be a living lab. This was for our university anyways the first lead platinum building was supposed to be the lead platinum building.

3:49 So there’s a few innovative things that get put into this building. I mean ultimately the building itself is is a a place for computer science, for AI, for robotics, but it’s also a place for convergent science. And so what happens is here are some of the renderings from HK of the space itself. But what happens in the process of me following along in this in this four-year journey is the idea that we have to somehow treat this as a new type of building asset, i.e.

4:18 A living lab. But again, nobody really understood what that meant. So at some point I got a little frustrated and I just said, “Okay, fine.” I had been working on this digital twin work a few years prior at at Ericup and I said, “Well, why don’t we actually finally, you know, produce this building as a digital twin having also started that campus twin, which was just sort of a a play on the side.

4:45 At the end of the day, the digital twin for us really is a combination of these things. It’s human building interaction and we’ll get back to that. It’s the idea that we can use our BIM models which are pretty high fidelity at this stage of the game. Connect to our building automation systems and enable simulation and automation and eventually enable hopefully some generative AI. So this is in a sense the diagram I used to sell the dean on the idea and we moved it forward.

5:16 What we ostensibly also had to do in the very beginning is we said look we are not going to go buy offtheshelf software from the vendors. We went about this the hard way quite frankly but we also proved in some ways that this is really quite repeatable. So my colleagues who actually did the real let’s say full stack development work which I cannot do myself built the viewer built the converter built the connections and this was actually prior to the building being built and let’s say connected so we built this environment now again the real core objective that we had frankly was to turn our building into an instrument of science right it is not technically a digital twin, right?

6:07 I cannot go into my system and turn the heat up on a professor I don’t like. Quite literally, that’s the reason why we can’t do that. There is no feedback loop that you would see in digital twins with Rolls-Royce or NASA or things like that. Quite frankly, what we built here was an access point for all of the data that this building does produce so that researchers of any kind, not just building building types, but researchers from AI to in fact occupational health and medical types can access the data that the building does produce.

6:45 This is in a sense our system architecture. What you I think what you really should notice is that we’re connecting to the Honeywell system. Is the eaten are the eaten people in the room? No, they were here yesterday. We’re actually working on connecting to all the eaten stuff right now. And most importantly up in the top is at the end of the day, it’s a research platform for all kinds of people, not just our disciplines in this room.

7:09 What it really is, it’s basically taking the BAS garbly right, which they intentionally make very very obtuse, creating a JSON format and creating ostensibly an environment which all of you would be fairly familiar with which is an API, a bunch of collabs or notebooks, Python integrations for understanding and visualizing the data. And quite important to me at least and to a lot of my colleagues is being able to see the data in situ in 3D.

7:41 I think it really does actually impact what we can learn from these things. So ostensibly our system is it’s a BIM model, building information model. It’s a connection to the building automation system. It’s most fundamentally an API. But this is the kicker. It’s all about the humans and how they influence these buildings. Because if anybody’s ever done an occupation an o post occupancy analysis of a building, we design these buildings, we we engineer these buildings, we try to hit these specific targets and we’re never right.

8:17 Never. I mean, unless somebody can say we were okay. We’re never right. And it’s because we’re always doing overestimation. We aren’t actually using machine learning. We are not doing anything that intelligent. And so ostensibly what we’re doing here is enabling people to make more sustainable buildings and more accurate. This is what it looks like. I’ll show you actually all the interfaces in a second, but basically we’ve ported our BIM information into the web.

8:41 So it’s in the cloud. It’s accessible to anybody. We’ve located every single sensor point. There’s about 3,500 plus. For the structural engineers in the room, I tried to get strain gauges on the steel while it was being built. It didn’t happen, but I did try very hard. And people can go in and literally access this data in real time. We built a series of interfaces for people to use.

9:05 There’s the architectural model, the mechanical model, the structural model, there’s a GIS view, but most importantly, quite literally, it’s the API because we didn’t know how people would want to access this data, what they wanted to do with it. It wasn’t just for architects by any means. It was for everybody. And so the API is really quite critical. The other thing that I think is really important to overstate really is that this is insitue research.

9:30 So this is not sanitized. Once people are occupying the building which it is now occupied, it actually influences all of these sensors and all of the data that we’re capturing and it actually is a much better test bed for people who want to develop AI tools and it is an AI school based on real life data sets. Okay. So, just because, I come from Hollywood. I’m not really from Hollywood.

9:56 I have to play you a movie. And let’s see. I have to find my There it is. I got this now. And hopefully you’ll hear it. And in the silence before dawn, there exists a threshold between what we know and what we have yet to discover. Silence before. Apologies for technical issue. Audio duplication in recording. It’s not you, it’s us. Stand as a threshold an instrument of science.

10:38 This is a vation. When it senses sound like a circuit pulses, this is a rhythm of knowledge. It listens. It is a living. It every passes with the knowledge within it. It lands the building reveal is its energy. The air currents coming out as digital actions of distant planets. The building energies by building energy systems like a silent river and are conly observed of distant planets and controlled inity between systems and machine.

11:30 We are not with the servers and the virtual we are partners in an inter humanely reach achieve the unknown through the digital twin and we search past reality’s edge with every touch we reach into the unknown through the digital twin and just we still past reality’s ecosystem of innovation It invites us to see beyond the walls of our world. This is more to reach us into the future.

12:04 It is an ecosystem and to share innovation with our hands. It invites us to see beyond the walls of our world is where human intuition into the future. Logic of the past with our hands as the architects of a smarter. This is where human intuition meets machine logic. Not as echoes of the past but as deep apostic to a space where data and life more sustained to our future and knowledge transcends the ordinary.

12:41 Join us in this step to a space where data don’t just life out intertwine and knowledge transcends the ordinary. Join us in this living land where we don’t just dream about tomorrow. We nailed it. Well, I don’t think I could repeat that sound effect ever again, but but I’ll try someday. So I do get to work with some amazing people at USC and there’s a very famous cinema school.

13:21 One of my close colleagues is actually a filmmaker and we’ve made some films together for other projects and and so he did this for me. I think it’s really important in all the work that I get to work on and do to credit the people that actually do the real work for me. These are my colleagues Nico Simon and Ozam. Some of you may know Ozam and Simon from Autodesk Research back in the day.

13:48 They’re they’re very close colleagues and Nico is part of tracks and is also at MIT. For the last four years they’ve been helping me build this system. They also have been in lots of meetings with actually all the other people along the bottom of this screen which is it does take a village frankly to build the building but it also takes an entirely different kind of village to build the digital twin for the building.

14:14 If anybody is interested there is the QR code it’ll take you to the login page and the only rule we have for using this data is that there is some kind of affiliation with somebody at USC a faculty member of any kind. So it doesn’t have to be me and in fact I would hope it would not be me for everything that we hope that it will help generate.

14:38 And so there I am. If you email living labus USC, it does come to me at the moment. But it’s also open to anybody that’s really interested in getting access to real building data to derive their own research. And thank you very much. All right. Thank you. I’m all down for the King Tubby remix of the videos coming through. Maximum maximum delay. To see the full recording of this and previous presentations, as well as information about future CDFAM events, visit CDFAM.com.