During our previous project between 2012-2017, we developed basic ethological, neurophysiological, and molecular genetic procedures that allow to better understand the mechanisms behind the socio-cognitive functions of dogs and humans. Based on our findings, we could broaden the scope of our research questions and apply some results in practice.
In our new project, applying synthetic ethology (SE), our aspiration is to collect a large amount of behavioral and neural data to generate general rules that allow the construction of new „in silico” agents. The SE approach is analogous to „synthetic” biology, primarily applied in molecular systems. SE is complemented by ethorobotics, an approach that aims at designing autonomous social robots capable of fulfilling a useful function living in a common group with other living (and/or similar robot) individuals engaging in social interactions.
Our basic aim is to develop a social interaction model based on the results of specific comparative cognitive tests, animal–robot interaction observations, and comparative neurobiological measurements. This synergistic approach allows us to create a novel and effective mindmodel.
1. Comparative analysis of human analogy of social behavior
As a continuation of our earlier research, we aim at further analysing the human-analogue social behavior observed in the dog, which is indispensable for the development of a comprehensive social behavior model. We investigate the behavior and behavioral synchronization of people and companions living in the same social context (host and dog) and compare their respective neurophysiological responses. Designing comparative studies also on family cats, we broaden the range of species included in the comparison to better understand the effect of domestication and individual developmental plasticity.
2. Examination of the mechanisms of social behavior through animal-robot interactions
The use of robots is a new direction in ethology in which our group played a pioneering role. In our previous studies, we developed the methodological basis for dog-robot interaction revealing that dogs may consider robots as a social partner to a certain extent after a short interaction. Our new question is whether a long-term, intensive, and strict rule-based interaction can create effective visual/vocal communication between dogs/cats and robots.
One technological basis of the interaction is to place a sensor on the subject that recognizes the behavior of its carrier on the basis of motion parameters. This information is also available to the robot so it can change its behavior depending on the behaviour of the dog/cat.
3. Study of the brain for the mental processes of communication
We have demonstrated that dogs process the human words in a specific way. Further fMRI studies could reveal how the dog, domesticated in a human environment and growing up in the family, processes certain linguistic precursors. The so-called sociocognitive precursors help interact with people (e.g., social attention, contingency reactivity), while computational precursors play a role in recognizing specific language patterns (e.g., syntax processing). Our question is whether, besides sociocognitive precursors, computational precursors are also processed in the brain of a non-primate species. In the case of specially trained dogs, we investigate relevant brain functions before and after learning about specific communicative signals.
4. Human-robot interaction
Our goal is to create a social robot that, in addition to the simplest, most robust construction, meets a certain function and can interact with people. Based on the data collected on dogs, we build a social robot that, with some limited language skills and using non-linguistic vocal signals, will become an effective collaborative helper for human partners.
Most of today’s social robots are not designed for “field” applications, so usability testing is hardly done. We test our social robot in real-life situations creating specific problem scenarios and examine people’s reactions (e.g., emotion expression, language, etc) applying ethological methodology (i.e., coding behaviour).
Minimum sociocognitive abilities.
What kind of sociocognitive skills should a social robot have to be able to interact autonomously with different living beings (dog, human)? Such a social model is a good starting point for robots to interact with people.
Non-linguistic basic elements of a complex communication function.
The effectiveness of a communication system is significantly affected by non-linguistic skills. We consider that social attention, turn-taking, synchronization, emotion-trapping, inhibition, and collaboration skills improve the communication of the two most popular species kept as companions (dog, cat) with humans.
Neural representations of human language.
Based on our recent findings, using comparative functional magnetic imaging (fMRI) and polysomnography (EEG), we measure dogs’ and humans’ brain representations using similar non-invasive methodology and compare their neural responses to different linguistic stimuli. Exploring these representations can help to better define and program the language skills needed for social robots developed for different purposes.
Most important pillars
Comparative behavioral research based on the dog.
In the modelling of social behavior, the dog has a central role in the human environment with very different genetic background (breeds) and phenotypes of social abilities (stray, family and working dogs) and environmental effects (training, keeping conditions, experience, etc.).
Comparative non-invasive neurobiology.
For the dog, we developed a novel methodology for non-invasive fMRI and electroencephalography (EEG) making dogs’ brain function comparable to that of humans in many respects. This allow us to search for a novel evolutionary framework by examining in which contexts the similarities/differences of the brain functions of these two species can be best explained by homologies or analogies.
Extending comparative research.
It is recognized that the socio-cognitive examination of the so far neglected domestic cat can bring a whole new aspect to comparative research. Even though cats have become one of the most popular companions of humans, they are claimed to be much less cooperative with their owners and try to exploit the benefits of the human environment to their full potential. However, systematic research is lacking.
Research of agent–human/animal interaction.
We found the basic methodological solutions that make communication and collaboration-based interactions measurable between lifeless and living agents. We argue that the interpretation of dog/cat–agent interactions can help construct social robots that successfully integrates into the modern human environment.
Applicability of our results in practice
Practical applicability is important to us.
Investigation of the dog/cat-human interaction improves the quality of life of companion animals. Our research is closely linked to the EU-sponsored “One Health” concept (Gibbs 2014), which essentially seeks to safeguard the health of humans and their companions in synergistic studies, including mental aspects.
- During the previous period we managed to create the hardware and software that provides the prototypes of our self-developed social robot. We, jointly with industrial and service partners, are planning to test robots in a live environment, in a real working environment (e.g., Enjoy Cafe Budapest).
- We believe that the integrated results of the behavioral, neural, and robot experiments (e.g., studies on dog/cat– robot interactions) can lead to novel insights that can be used both for diagnostic and therapeutic purposes, for example, in case of children with ADHD.