Interdisciplinary research, data ownership and open science

The artificial culture project team is starting to place data on its website and make it publically available – well some of it – the data from robot experiments. The data so far uploaded is text so file size is relatively small.

For developing the open science aspect of the project, along with Sajida Bhamjee, I am working out how to make available video of the robot experiments. This has raised technical issues that prompt me to reflect about data ownership.

To maintain the project data availability over many years, perhaps well after the current research team have moved on, we have chosen to use freely available web space from one of the major commercial suppliers, so have to work within the limitations of this. We could buy more space but we do not have a funding stream to support this into the future. We can set up web space in the name of the whole team with one quota of space, or each team member can set up personal spaces and thus gain access to a larger quota of space. In a team space with equity of access, everyone can change everything. In individual spaces the person running the space specifies what is shared with other individuals or open to the public, and specifies who can change items in that space. Using individual space for the project means as individuals, we are using space we might have used for other things.

When uploading onto You Tube, the film of the social scientists meeting the robots, filmed summer 2010, it was the first time I had done this and so I just followed the instructions and opened an individual account. When I then went to upload onto You Tube a video of robots imitating, it occurred to me that, apart from the intellectual contribution through project team meetings, all the work to produce the film had been by other team members. I felt uncomfortable about uploading it under my name. We went ahead as we were exploring technical aspects of sharing videos. Realising You Tube was not the best place for the video (as it was difficult for people to download it and this was important as the video needed to be on fast forward to be watchable) we turned to using google docs. However, we face the same problem: into who’s web space do we place the video?

When sharing data within a team where members undertake similar research – for me this is research using social science methods such as interviews – data is shared by the team even when only one or two team members actually collected the data. The rest of the team undertakes tasks to enable the data collection including gaining the funding, enabling access for recruitment of people to the study and supervising the data collection, and brings analysis expertise to the data. For projects where I didn’t collect the data, I know that given sufficient time (by which I mean time out from other tasks), I could collect the data myself - I have the skills to do this.

One of my uncomfortable feelings about uploading the video of the robots onto my You Tube website, was that I could not collect the data myself. It is just possible that given sufficient time to retrain as a roboticist I just might – but this would be a whole career not time out from teaching and administration. I can contribute to analysis of the data, but not in a way that would be sufficiently robust for critical review. It would be easy for me to interpret something about the robots that I thought was interesting. However a swarm roboticist might realise that this was a technological hitch interfering with our experiment rather than a result of the experiment.

Investigating social networks

How do we operationalise the notion of social networks for the robots in the artificial culture project?

In the social world embedded in the biological and physical world, interaction occurs between individual people (agents), between organisations or groups of people acting almost as if a single agents, and between individual people and organisations/groups. I am deliberatly being vague about what consitutes an organisation or group – it could be a clique of friends or a nation of people – there are questions about the location of boundaries between organisations or groups that I am not tackling here, nor the issue that individual people make up the organisations and groups.

Messages (information, ideas, stories, advice etc) move between agents – individuals as agents and organisations/groups as agents. The locations and modes of interaction are very varied and include broadcast media, face to face interaction, interaction via the internet. There are different constraints and different potentials for what is exchanged, between whom and the timing of any exchanges – synchronous or asynchronous and by how much.

The messages and the agents are often to some extent distinct. It could be argued that there will be emergence from message interaction, which might be culture or norms, and emergence from agent interaction, what might be considered as social structures, and from the interaction of messages and agents. Agents change messages and messages change agents. Agents and messages also influence the mode of interaction. Society is complex and trying to unpack it is difficult. With the robots, we are starting with individual robots and attempting to enable the development of a simple version of this.

In the artificial culture we might consider the robots as the agents. Currently they are programmed to look for a signal from another robot, watch the robot’s dance (pattern of movement) then, after signaling, imitate what the robot observed. Meanwhile other robots are looking for a signal and then observing to then signal and imitate. The message that is transmitted from one robot to another is the dance. So we have an interaction – that of observation and imitation of each other.

One link in a human social network is usually defined qualitatively in terms such as friends or best friends or more quantitatively in terms of amount of social interaction such as talking to someone, with a measure of frequency the interaction. Many of the studies on social networks, at least those related to health (my area of interest), are cross sectional studies and so not concerned with formation of social networks. Those that are longitudinal still start from an existing social network. So, do we programme a social network into the robot swarm or do we programme a behaviour that might lead to the development of a social network?

We could design a social network consisting of robots with a propensity to more quickly notice the signaling of the robots they are linked to, compared to robots to which they are not linked. I think this could be achieved through radio signals.

An alternative is to programme each robot to notice which robots do a dance that is most similar to their last dance (within a certain duration) and to then notice this signaling of this robot more quickly than other robots.

Doubtless this needs more refinement.

Tracks, roads, footpaths and social networking

As we develop our swarm of interacting robots, should we take consider introducing structures within our arena that are analogous to those at the micro-biological level inside cells? Packets of molecules move around cells along tracks or microtubules carried by molecular motors. See http://mechanochemistry.org/movies/ (the movie called Golgi H264-33 shows the tracks or microvilli with the packets of molecules moving along them). These tracks are influenced by what is carried along them and refashion themselves accordingly.

We can draw analogy with people - tracks, roads and footpaths that people move along, and perhaps with social networking where it is messages that move along.

The tracks take the molecule, the person or the message to the place where and interaction occurs - with another molecule, another person or the message is received by a person. At present, our robots can move in any direction within their arena. Should we consider having some tracks and meeting places?

Robot imitation experiments with children

I have spent much of the last few weeks pondering about the different ways in which I can show children the e-pucks imitating each other. I propose one of three different methods:

(1) Show the imitation video, ask the children what they think is happening and then show them the player stage animation film stating ‘this is what ‘actually’ happened’.

(2) Show the imitation video and the player stage animation film side by side.

(3) Show the imitation video alone.

However, I have concerns with all three methods. With the imitation videos alone, it is very hard to see the patterns (even when I know what I’m looking for!) From past experience, I have learnt that children will most likely give a response to a question regardless of whether they know the answer or not. Therefore will they just provide a response to please me? Conversely, if I show the children the player stage animation films, will I be ‘spoon feeding’ them with responses? Will I be making the patterns too apparent?. In addition, by demonstrating the imitation video next to the player stage animation film the children’s attention may be divided.

Another problem with robot videos is that it may not be as engaging as having embodied robots in front of the children. I suppose one might ask ‘Why don’t you take the robots to the children?’ Unfortunately, it is not a simple process of switching the robots ‘on’ and ‘off’. There are various tracking and logging pieces equipment that is only available at the robot lab.

After having a lengthy conversation with Andy (a fellow team member), he suggested that it may be possible for the robots to leave a ‘trail’ as they did the imitation. The benefit of that method instead of method number 2 (showing player stage animation film and imitation videos simultaneously) is that the children’s attention is less likely to be divided. This still however leaves the question of ‘am I making the patterns too obvious?’ The highlighted drawings (player stage animation) will clearly show that the robots are producing shapes. However, it is questionable as to whether children will associate this with imitation.

Robot interaction seen through the eyes of children

Our project hopes to discern new patterns of behaviour emerging from robot interaction. A concern is that the robot team will be unable to discern new patterns – ‘unable to see the wood for the trees’. Might it be that children are able to see patterns that as the research team we would miss? One of our team members, Sajida Bhamjee, has taken e-pucks into schools and collected data on the children’s reactions to seeming them moving around. The e-pucks were programmed to follow each other by sensing each other’s tail light and following it. Due to variations in the environment, such as light levels, there was some variation in what they actually did. The e-pucks moved quickly and bumped each other. The children talked in interesting ways about what was happening and who was controlling it.

Our next step is to show children the robots imitating each other. It is difficult to set this up outside the robotics lab (not impossible but time consuming and technically challenging), so instead of taking the robots to the children we have filmed the robots. However, when they imitate they move slowly as you can see if you follow this link to the film on You Tube:http://www.youtube.com/watch?v=hygWbKcAaTs

This might not interest children. Would it catch their attention if speeded up? The video is available for down load. Please try viewing the film speeded up.

From the data collected through the tracking equipment in the robot lab, and using ‘Player Stage’ software, the robots as seen in the film can appear as an animation with their tracks as if ‘drawn in the sand’. You can download a video of this. Below is the tracks they drew.

It maybe that the children would enjoy seeing the animation more than the film of the robots. There is also the question of whether the animation should show the 'tracks in the sand' or not.

Open Science and the Artificial Culture Project

From the start of our project, the intention has been to make it open to all who are interested, and to enable this by having an interactive project website and to put our data on the website. The idea of open science was new to me but immediately seemed attractive. Why? Here are some reasons I find persuasive. Open science opens up opportunities for ideas to flow into the project from anyone who is interested. It increases access to the process of doing science to the public and makes us as scientists more directly accountable to the public. It allows others to analyse our data to check our analysis or try new ways of doing the analysis and there is potential for others to discern emergent patterns in the data or maybe the research process, that we as a team do not notice as we are part of the project.

So why have I found it hard to write about the project on our project blog? There are technical reasons such as forgetting the password for signing in and not being online when I have an idea for a blog entry. However, a deeper reason is feeling I do not have anything worth saying – or at least not yet.

My ‘not yet’ includes a whole mixture of assumptions and feelings of which these are a few have clarified through discussion with Ann Grand and Alan Winfield.

The process of doing research is not interesting, only the result, and even then the result should have some obvious implications for the global community and environment.

Actually I don’t believe that. As a sociologist my research includes the study of processes of doing and using science. Science involves a lot of people working away at small parts of a problem, without being sure of the implications of what they are doing. If we knew the results of our research beforehand it would not be research. We can of course suggest why it might be important to investigate a particular aspect of the world, but we cannot know for sure that our research will have impact.

Research is about finding new knowledge so it I need to check something is new before I talk about it.

To be sure of this takes a lot of time as it involves searching the literature and reading everything that is relevant. This approach dominates health sciences and arguably is the appropriate approach in this field of research, which is applied to improving health. However, from working with scientists from other disciplines, I understand that there is not this obsession with searching for all relevant literature. The difference seems to be that in health sciences we are often working from a current understanding of a biological, psychological or social mechanism and then testing whether an intervention, based on this understanding, then makes a difference when used in a population. In biological, and indeed social science, we are more often seeking to understand the mechanism. This demands a different critical approach to what is already known.

But there are also feelings..

So far my reasons for ‘not yet’ are bound up with the research disciplines in which I work – sociology and health sciences. However there are also my feelings. There is the fear that people will judge me negatively and the fear that people will take my ideas and beat me to the credit. These fears seem to be contradictory, suggesting that I feel I am both a terrible scientist and a brilliant scientist. Having contradictory feelings is very normal. However, I think these fears are bound up with the results of research rather than the process. I know it is possible to undertake research that has a rigorous process but does not make a world-changing discovery. I spend time teaching rigorous research processes to students. So, although the feelings are unlikely to go away, I can perhaps muster the confidence to open up the research process to scrutiny. Anyway, if I do have a good idea and then someone else gets the credit, I will be in the good company of all the other scientists who have put in place the building blocks for the world-changing discovery.

Robot imitation as a method for modelling the foundations of social life

Robot imitation as a method for modelling the foundations of social life: a meeting of robotics and sociology to explore the spread of behaviours through mimesis

Here is the video, posted earlier this month by Frances Griffiths on YouTube, of the meeting of robotics and sociology I blogged about on 21st June. No need for me to write anything more - Roger Stotesbury's excellent 10 minute film explains the whole thing...

Open-ended Memetic Evolution, or is it?

Just finished a paper describing some new results on open-ended memetic evolution from the Artificial Culture project. I describe in some detail one particular experiment in which 2 robots imitate each others' movements. However, here the robots don't simply imitate the last thing they saw; instead they learn and save every observed movement sequence, then when it's a robot's turn to dance it selects one of its 'learned' dances, from memory, at random.

Here is a plot of the movements of the 2 robots for one particular experiment; this picture has been generated by a tool developed by Wenguo Liu that allows us to 'play back' the tracking data recorded by the Vicon position tracking system. The visualisation tool changes the colour of each 'dance', which makes it much easier to then analyse what's going on during the experiment.


Epuck 9 (on the left) starts by making a 3 sided 'triangle' dance, numbered 1 above. Epuck 12 (on the right) then imitates this - badly - as meme number 2, which is a kind of figure-of-8 pattern. It is interesting to see that this 4-sided figure-of-8 movement pattern then appears to become dominant, perhaps because of the initially poor fidelity imitation (1 → 2), then the high fidelity imitation of 2 by epuck9 (2 → 3), then the re-enaction of meme 2 as meme 4. And then subsequent copies of the same figure-of-8 meme then appear to be reasonably good copies, which reinforces the dominance of that meme.

Since the robots are selecting which observed and learned meme to enact, at random, then there is no 'direction' to the meme evolution here. Memes can get longer or shorter - both in the number of sides to the movement pattern, and the length of those sides, and the resulting patterns arise in an unpredictable way from the imperfect 'embodied' imitation of the robots. Thus, we appear to have demonstrated here, open-ended memetic evolution.

Warwick Mimesis project visit to the lab

As a follow-up to a talk I gave last December in Warwick, we were visited in the lab today by a group of social and complexity scientists from Warwick including Frances Griffiths, Steve Fuller and Nick Lee. We had a hugely interesting day discussing the extent to which (or, indeed, if at all) robots could be used to model mimesis in society.

The day started with me describing the embodied imitation-of-movement experiments that we are currently doing here within the Artificial Culture project, and demonstrating the latest version of the Copybots experiment. After lunch we then had a round table discussion about whether or not such a simple model might have value in social science research and - somewhat to my surprise - there seemed to be strong consensus that there is value and that this (radical) new approach to embodied modelling is something we should actively pursue in future joint projects.

The meeting was filmed by Roger Stotesbury of Jump Off The Screen and I hope to post a link to the video record of the meeting on this blog.

Postscript: here is my blog post with Roger's film of the meeting.

EPSRC HOW? event

Spent a most interesting day today at EPSRC HQ in Swindon. I was one of several academics asked to come and exhibit their work to the staff of the EPSRC. The idea of the event was to enable all of the staff of the council to get an insight into the research that EPSRC funds when, in the normal course of events (I guess), only a relatively few would get to see that research - programme managers for instance.

I took along some e-pucks and a portable arena, which proved very popular, together with this poster for the Artificial Culture project.