Wevolver Robots in Depth

Gabriel Skantze talks about how he works with human robot communication, and about how the social robot revolution makes it necessary to communicate with humans in a human ways through speech and facial expressions. This is necessary as we expand the number of people that interact with robots as well as the types of interaction.

Gabriel gives us more insight into the many challenges of implementing spoken communication for co-bots, where robots and humans work closely together. They need to communicate about the world, the objects in it and how to handle them.

We also get to hear how having an embodied system using the Furhat robot head helps the interaction between humans and the system.
Having an expressive face like the Furhat adds many improvements to how a system can communicate with people. It also improves the human engagement and understanding of what the system tries to communicate significantly.

Gabriel then talks about the how they use AI and machine learning to understand speech. Understanding an individual speaker’s way to speak, thus adapting a robot to its user can improve their communication.

As the Furhat system is used out in the field, we get valuable insights from real world situations. One such case is guiding travellers at an airport to improve their experience and make travelling more efficient for everyone.

This podcast is part of the Wevolver network. Wevolver is a platform & community providing engineers informative content to help them innovate.
Learn more at Wevolver.com

Lars Dalgaard shares his experiences developing robots in many different contexts.

We hear about Lars’ early work with large mobile robotics in a commercial nursery garden handling the transfer of plants from the greenhouse to the field.

He also speaks about the Hydra project a self re-configuring modular robotics project that developed several different modular robotics systems including the Atron system.

Lars felt that there was a problem with the process used to introduce robotics and automation into society. Commercialization was hard and unreliable mostly because there was no focus on designing a complete system. This lead to an industrial PhD done at the Danish Technological Institute (DTI) focusing on a system level design approach.

He then talks about his work at DTI that focuses on transferring research, knowledge and research results from academia into companies and the Danish society in general.

One project Lars has been working on is augmenting mobile platforms so that they can handle tasks as they move around in a production facility. This also aims to make it easier to program the mobile platforms and any systems added to them.

Lars thinks that looking at the bigger picture and bringing multiple partners and end users into projects, and doing so early, can bring big benefits to a project.

This podcast is part of the Wevolver network. Wevolver is a platform & community providing engineers informative content to help them innovate.
Learn more at Wevolver.com

Harri talks about AI and how he aims to mimic human decision making with algorithms.

Harri has done a lot of AI for computer games to create opponents that are entertaining to play against. It is easy to develop a very bad or a very good opponent, but designing an opponent that behaves like a human, is entertaining to play against and that you can beat is quite hard. He talks about how AI in computer games is a very important story telling tool and an important part of making a game entertaining to play.

This work led him into other parts of the AI field. Harri thinks that we sometimes have a problem separating what is real from what is the type of story telling he knows from gaming AI. He calls for critical analysis of AI and says that data has to be used to verify AI decisions and results.

We also hear about the current use of AI in among other things sports games, where the challenge is to make a believable computer copy of a real-world player to make the games feel more real.

We then get to hear about his current work in developing AI systems that create mind-maps from texts to make the computer able to determine the context. By doing this we can derive better sentiment and meaning. This AI system is trained using a large amount of reliable data from scientific papers and CNN.

One application of this that Harri is working on, is in the labor market matching needs for talent with available people. This will help companies find staff and people find job opportunities.

This podcast is part of the Wevolver network. Wevolver is a platform & community providing engineers informative content to help them innovate.
Learn more at Wevolver.com

Promote your company in our podcast?

If you are interested in sponsoring the podcast, you can contact us at [email protected]

Christian talks about AI and wanting to understand intelligence enough to recreate it.

Christian discusses building systems that can interact with humans beyond regular computer interfaces.

He started working with computers early, too early for advanced AI. He studied broadly, including philosophy, ethics as well as synthetic and biological neural networks.

Christian has found a lot of inspiration in the old papers by Alan Turing and others. He sometimes envies them their opportunity to think big, as they did when they founded the areas of computer science and artificial intelligence.

Christian has be focusing on AI in healthcare and has recently started to communicate the opportunities and challenges in artificial intelligence to the general public. This is something that the host Per Sjöborg is also very passionate about.

We also get to hear about the Nordic AI institute (https://www.nordicaiinstitute.com) and the work it does to inform all parts of society about AI. Anyone interested in AI is welcome to reach out if they have questions or if they have knowledge to share.

Then we hear how Christian is working with Tieto (https://www.tieto.com) on integrating artificial intelligence in their customers daily operations.

This podcast is part of the Wevolver network. Wevolver is a platform & community providing engineers informative content to help them innovate.

Learn more at Wevolver.com

Promote your company in our podcast?

If you are interested in sponsoring the podcast, you can contact us at [email protected]

Julien Bourgeois talks about self-reconfiguring modular robotics and how he is developing millimeter sized robots called Claytronics.

Julien started out as a computer scientist. He was always interested in robotics privately but then had the opportunity to get into micro robots when his lab was merged into the FEMTO-ST Institute. He later worked with Seth Copen Goldstein at Carnegie Mellon on the Claytronics project.

He tells us how he works on creating a world built with programmable material that would allow objects to change their form and function automatically by running a program. This will create smart objects that can adapt to the world around them and user preference in a totally new way.

One large benefit of programmable matter is that development can happen both in the computer and in the real world with changes transferred between them. A change done in the code would appear in the part made up of programmable matter, but the part can also be changed in the real world and the change would be transferred to the program controlling it. This would create a very flexible, dynamic and highly intuitive design process.

The structures based on programmable matter also exhibit many very special characteristics. They can be self-healing if they get damaged, they can dynamically respond to load and be as strong as needed, they can degrade gracefully and predictably and can even indicate that they are overloaded and might fail so that the user can take the appropriate actions.

We also learn about a system for sorting very small components he built and how cameras could not be applied.

Per and Julien discuss how developing programmable material is hard and that many difficult problems have to be overcome. At the same time, many problems with the current way of doing things will be solved in a fundamentally better way by systems built with
programmable matter.

Julien shows an enlarged mock-up of the small robots that make up programmable matter, catoms, and speaks about how they are designed. Currently he is working on a unit that is one centimeter in diameter and he shows us the very small CPU that goes into that model.

There is also an art project in progress, using another version of programmable material building blocks.

More about the small CPUs mentioned at https://www.cubeworks.io This episode was recorded at ICML, IJCAI-ECAI, AAMAS in Stockholm, Sweden 2018.

This podcast is part of the Wevolver network. Wevolver is a platform & community providing engineers informative content to help them innovate.
Learn more at Wevolver.com

Promote your company in our podcast?

If you are interested in sponsoring the podcast, you can contact us at [email protected]