Frequently Asked Questions
What’s in it for me?
Our aim is to foster interdisciplinary connections and potential future research collaborations among diverse scholars. Participating in our workshop will allow you to network with others who share the same interest, while also enjoying a drink. We will guide you through networking techniques and research-focused exercises to break the ice and help you to build connections with others.
Without you, this dataset may not be possible! Once it is created, processed and ready, we will share it with the research community under a restrictive End User License Agreement for academic research. That means you can also use it for academic research.
What type of information do you extract from the multi-modal data?
We will manually or automatically annotate the data for various social behaviors, speaking status, redacted transcripts, social involvement, perceived intentions, conversational events, body skeletons, vocal activity, and potentially others in part or all of the data.
These annotations will be shared as part of the resulting dataset. The annotations will be used to train or finetune machine learning models to automatically extract this data. This processed data will also be shared.
By sharing the dataset with the broader community, other researchers may annotate the data for other behavioral phenomena that is closely related to social interaction. And also train or fine-tune their own Machine Learning models. Anyone using the data will be required according to the End User License Agreement to share these research results with the community.
Can I participate in the workshop without contributing my survey data or being recorded?
Unfortunately not. Our workshop room will be equipped with multiple overhead cameras that record the whole event and the individual cameras will record audio. All participants must agree to the full collection of all data. However, if you are interested in the work we do, feel free to reach out and connect with us outside of the workshop!
Who will the data be shared with?
The dataset can be shared (subject to a signed End User License Agreement) with any researcher at a legitimate organization that carries out academic research. To apply and have the dataset granted to you, you must be at least an Assistant Professor or equivalent grade. Access to more junior supervisees is possible; the supervisor agrees to be responsible for the correct handling of the data by the supervisees. All applicants will be vetted before the data is released to them. If applicants to not adhere to the requirements for application, their request will be rejected.
Will my data be anonymized?
Depending on the type, your will be shared in either a pseudonymized or anonymized form. This means that your pseudonymized and anonymized data is linked to a anonymized numerical identifier. This identifier cannot be traced back to your name and identity.
Your video and audio data is pseudonymized. This means that your video and audio data could be deanonymized by intentional abuse. To handle this, the End User License Agreement stipulates that any form of deanonymization of the raw data is strictly forbidden. In the unlikely event of a breach, Delft University of Technology will take legal action against the individual(s) in question.
I don’t work at a University. Can I take part?
Yes! However, our ethics board do not allow us to share data for commercial or governmental research. The EULA only allows for academic research to be carried out with the data.
Can I request the removal of my data after my participation?
As stipulated in the consent form, this is possible for a fixed period after the end of the event. We will delete all data. In the case of the videos, we will you redacted from the data using automated tracking tools. We would however, aim for this not to happen. So if you plan to participate with the intention of asking to have the data deleted, we request you to refrain from doing so. The more complete the data is, the more impactful for the community.What if I signed up to participate, but cannot make it to the workshop because my schedule changed?
If you signed up and we have confirmed your registration, but you then find out that you cannot make it to the workshop – please let us know as soon as possible. The way our workshop is designed requires an exact number of participants. So, the sooner we know that you cannot attend, the sooner your spot can be filled by another interested scholar. J
When will you share the dataset?
Post processing, annotating, and training all the models for this data requires some new research to be carried out. We aim for this to be completed and ready within 12-18 months with staggered release where for more straight forward data will be released sooner and more complex will be released later.
This sounds super interesting and I want to collaborate – what should I do?
Come to our workshop! J If you cannot make it, feel free to reach out to us and we can connect outside of the workshop. You can find our contact information here (link to contact sites). Concretely, if you are willing to help with some of the tasks involving post-processing the data, that would be great.
I feel inspired! How can I run my own version of ConfLab?
The entire set up is available either open source or with commercially available resources. We strive for reproducible documentation including sharing share code and pre-trained models. A guide was published at NeurIPS . Feel free to contact us if you have any questions.