Natural Language Generation for Nature Conservation: Automating Feedback to Help Volunteers Identify Bumblebee Species

Proceedings of COLING 2012: Technical Papers, pages 311–324, COLING 2012, Mumbai, December 2012. Natural Language Generation for Nature Conservation: Automating Feedback to help Volunteers identify Bumblebee Species Steven Blake1 Advaith Siddharthan1 Hien N guyen1 Nirwan Sharma1 Anne-Marie Robinson2 Elaine O′Mahony3 Ben Darvill3 Chris Mellish1 René van der Wal2 (1) Department of Computing Science, University of Aberdeen, U.K. (2) Aberdeen Centre for Environmental Sustainability (ACES), University of Aberdeen, U.K. (3) Bumblebee Conservation Trust, University of Stirling, U.K. s.blake.08@aberdeen.ac.uk, advaith@abdn.ac.uk, h.nguyen@abdn.ac.uk, n.sharma@abdn.ac.uk, annierobinson@abdn.ac.uk, elaine.omahony@bumblebeeconservation.org, ben.darvill@bumblebeeconservation.org, c.mellish@abdn.ac.uk, r.vanderwal@abdn.ac.uk ABSTRACT This paper explores the use of Natural Language Generation (NLG) for facilitating the provision of feedback to citizen scientists in the context of a nature conservation programme, BEEWATCH. BEEWATCH aims to capture the distribution of bumblebees, an ecologically and economically important species group in decline, across the UK and beyond. The NLG module described here uses comparisons of visual features of bumblebee species as well as contextual information to improve the citizen scientists’ identification skills and to keep them motivated. We report studies that show a positive effect of NLG feedback on accuracy of bumblebee identification and on volunteer retention, along with a positive appraisal of the generated feedback. KEYWORDS: NLG, Natural Language Generation, Educational Application, Bumblebee Conser- vation, Citizen Science, Generating Feedback. 311 1 Introduction There is a growing realisation of the potential of digital approaches, including the use of websites and social media, to increase participation in “citizen science”, which includes observing and monitoring the natural world. For instance, in the UK, the Open Air Laboratories (OPAL) network (www.opalexplorenature.org) is a large current initiative led by Imperial College, which aims to create and inspire a new generation of nature-lovers by getting people to explore, enjoy and protect their local environment (Silvertown, 2009). Within OPAL, iSpot (www.ispot.org.uk) is an online nature community that connects beginners with experts and fellow enthusiasts. Other groups have explored the use of standard social networking sites to generate public interest and collect data about the distribution of species (Stafford et al., 2010). Publicly available resources include www.thewildlab.org, which provides software for a number of mobile platforms, and www.scienceforcitizens.net, which acts as a forum for citizen scientists to find out about projects they can participate in and for researchers to publicise their projects. Although digital tools can enthuse the public and be used to enlist (for a short time at least) willing volunteers for nature conservation projects, initiatives such as the above have to contend with at least the following issues: 1. Data quality: participants are generally untrained, and not necessarily motivated to produce high-quality data (Stafford et al., 2010). This is not a problem if a project is primarily an “outreach” activity, but many projects also have specific scientific goals. 2. Retention of volunteers: in order to secure continuing participation, systems need to constantly renew to keep users interested and engaged. Existing digital support for nature conservation volunteers can only give them feedback and encouragement by allowing them to interact with human experts or by showing them pre- …

Content truncated to 500 words