About BeeWatch

Why are we doing this?

We are investigating whether new computer science technologies can benefit nature conservation programmes. Computer scientists and ecologists from the Aberdeen University are working with the Bumblebee Conservation Trust on the online identification tool `BeeWatch'.

Two avant-garde computer science technologies are being used in BeeWatch: Natural Language Generation and Crowd Sourcing. Both technologies help you to become familiar with bumblebee identification, and generate a better collective understanding of the distribution of these important pollinators within our country.

Natural Language Generation (NLG) is a technology that analyses data to generate text. In BeeWatch NLG is used to generate feedback to participants. Immediately on submission, submitters who offer an identification of the bumblebee species receive feedback that contextualises the record with respect to historical data held by BeeWatch and the NBN. BeeWatch will then verify the species and come back to the submitter by email. The email text is also composed by a computer, but edited by an expert where needed. When BeeWatch determines the bumblebee to be a different species than indicated by the submitter, differences in identification features between the two bumblebees are used as data that form the basis of a text. This text is subsequently organised, screened against linguistic constraints and finally embedded in a wider feedback message which the expert uses as basis of the feedback emailed to the submitter. This allows our experts to concentrate on the identification and enables feedback that is much richer than otherwise would be the case due to time and resource constraints. We are continuing development of the NLG, and you can expect to see more contextualised feedback soon.

Crowd sourcing is the practice of obtaining information from a large group of people, often on-line. We are developing this practice for small groups of people with an interest in bumblebees, which one could call Group sourcing. BeeWatch needs such an approach because more and more people are submitting photos and we want to scale this up even further to obtain as much information as possible on the geographic spread of these important pollinators. Rather than having a few experts attempting to identify thousands of photos, we ask other people to suggest identifications. For many bumblebee species this leads to clear ‘majority votes’ in which case there is little use for an expert. In other cases, opinions are divided and an expert will conduct the identification.

The development of NLG feedback (and the infrastructure around it) has helped experts to identify bumblebees for far more submitted photos than in the past. Group sourcing allows many more people to take part in BeeWatch and train themselves up in bumblebee identification. If enough people take part, BeeWatch will be an unprecedented tool that allows society to see current distributions of all our bumblebee species and changes over time.

Where can I learn more about the science?

  • Rene van der Wal, Nirwan Sharma, Anne-Marie Robinson, Chris Mellish and Advaith Siddharthan. 2016. The Role of Automated Feedback in Training and Retaining Conservation Volunteers: a Case Study of Bumblebee Recording. Conservation Biology, John Wiley & Sons.
    Paper (pdf) / Show Abstract

  • Advaith Siddharthan, Christopher Lambin, Anne-Marie Robinson, Nirwan Sharma, Richard Comont, Elaine O'Mahony, Chris Mellish and Rene van der Wal. 2016. Crowdsourcing without a crowd: Reliable online species identification using Bayesian models to minimise crowd size. ACM Transactions on Intelligent Systems and Technology (TIST).
    Paper (pdf) / Show Abstract

  • Rene van der Wal, Helen Anderson, Ane-Marie Robinson, Nirwan Sharma, Chris Mellish, Ben Darvill, Advaith Siddharthan. 2015. Mapping species distributions: comparing the spread of UK bumblebees as recorded by the national repository and a citizen science approach. AMBIO: A Journal of the Human Environment, Springer Science.
    Paper (pdf) / Article at Springer.com / Show Abstract

  • Steven Blake, Advaith Siddharthan, Hien Nguyen, Nirwan Sharma, Anne-Marie Robinson, Elaine O'Mahony, Ben Darvill and Chris Mellish. Natural Language Generation for Nature Conservation: Automating Feedback to help Volunteers identify Bumblebee Species. In the Proceedings of the 24th International Conference on Computational Linguistics (COLING 2012), Mumbai, India.
    Paper (pdf) / Show Abstract



  • Who are we?

    BeeWatch is co-ordinated by researchers at the University of Aberdeen in collaboration with the Bumblebee Conservation Trust (BBCT). The new Planting for Pollinators tools on BeeWatch have been developed in collaboration with BBCT, the Royal Horticultural Society (RHS), BWARS (Bees, Wasps and Ants Recording society) and Open Air Laboratories (OPAL).

    Beewatch was supported by RCUK's Digital Economy theme as part of Aberdeen University’s Digital Economy Hub to explore how digital technologies can help communities transform the way they manage, use and conserve natural resources.

    Planting for Pollinators is supported by the Natural Environment Research Council (NERC).

     

    Bumblebee Conservation Trust is a charity that supports the conservation of bumblebees, and raises awareness and understanding of bumblebees.

     

    The Open Air Laboratories (OPAL) network, led by Imperial College London, is a UK-wide partnership initiative that inspires communities to discover, enjoy and protect their local environments through citizen science-based activities. OPAL began in 2007 and is funded by the Big Lottery Fund’s Supporting UK-wide Great Ideas programme. For more information, please visit www.opalexplorenature.org or follow OPAL on Twitter @OPALNature.

     

    The Royal Horticultural Society (RHS) is established to share the best in gardening throughout the UK.

     

    BWARS is the national society dedicated to studying and recording bees,wasps & ants in Britain & Ireland.