Catching the User - Logging the Information Retrieval Dialogue Paul Landwich Claus-Peter Klas Matthias Hemmje FernUniversität in Hagen FernUniversität in Hagen FernUniversität in Hagen Germany Germany Germany paul.landwich@fernuni-hagen.de claus-peter.klas@fernuni-hagen.de matthias.hemmje@fernuni-hagen.de ABSTRACT As stated above we need to log all user and system ac- This position paper supports the idea of the information tivities and the corresponding result sets within a task to dialog between IR systems and users during an informa- catch the users context. From the experience one knows tion search task. In order to satisfy the communication and that a search task is usually not concluded with the first interaction needs of humans, IR systems should explicitly query. Rather a working context through the interaction is support the cognitive abilities of the users. An information elaborated. When this understanding becomes clear, there dialogue which does not only support an individual query must be some kind of accompanying information dialogue. but also the complete search process is necessary. Only in A dialogue consists of a sequence of activities and results. this way it is possible to satisfy an information need. In the past initial research ([2] and [6]) focused on the human users not only as a part of the system but also as an important component. In later works it was recognized Categories and Subject Descriptors that the search is a process. In other papers (e.g. [1]) the H.3.3 [Information Search and Retrieval]: Search Pro- search strategies and search patterns where investigated. cess; H.3.7 [Digital Libraries]: User Issues The overall complexity of the search process was exposed ([11], [12],[17]). In [7] a continuation models of information dialogue was introduced, to cover this search process. General Terms The process of the related research was consistent: Start- information retrieval, visualization, interactive systems ing from the support and improvement of individual queries, up to a more global view of the search process and dialogue. 1. INTRODUCTION But this global view must become granular again. In order to interpret a process or a dialog, the individual steps must Information seeking is usually not a single step to recover a be identified and formalized within this dialog. piece of information, but a cyclic, highly interactive process [8] identified six activities – exploration, navigation, focus, with the aim to satisfy a specific information need. Within inspection, evaluation and store – to focus on to derive a such a process the user builds a cognitive model, which helps context model of the user. her to reflect and advance the search process.Within user interfaces, it is necessary to integrate tools and functionali- 1. Exploration: The access to set of information ob- ties within existing tools, in order to develop this cognitive jects in the form of a query and the visualization and perception and derive a context model of the users. Require- realisation of the produced result set defines the Ex- ments for this are logging of all user and system activities ploration. A change respectively an enlargement of ranging from entered queries to the result sets, tools to vi- the informal context is caused by it. sualize the context and system support based on a context analysis. 2. Focus: The focus set represents the subset of infor- mation objects of a result set which reach the field of 2. ASPECTS vision of the user through a visualisation and is the result of the activity Focus In order to support the statement of the introduction, we would like to dwell on three aspects. 3. Navigation: The movement within a set of informa- tion objects (information room) or between different 2.1 Logging information rooms. This causes a change of the focus. 4. Inspection: Inspection is used for the cognitive de- termination of the state of an information object. 5. Evaluation: Evaluation gives the system a feed- back of the user’s understanding of relevance and ap- points the verified recall set. SIGIR 2009 July 19-23, 2009, Boston, USA. 6. Store: This activity allows to store found documents. Copyright 2009 ; Copyright is held by the author/owner(s). It either happens logically in form of a storage box on the user interface or physically when a document is C: R: Content set Relevance set Projection plan kv,total kf,total downloaded or printed. r1 ,r2,r3: Result sets C ki,3 ke,3 kf,3 kf,2 k rtotal Based on these definitions we can log a dialog or the kv,2 e,2 whole search process with the system. Because some of ke,1 C kv,1 kf,1 these acitivities correlate we can identify three interactive C r2 R r3 modes. The user finds oneself in one of theses modes and will ke,total r1 R change cicular the mode. The first mode is every time access. R Within this mode there is only one activity, Exploration. C R Already after the first Exploration the user changes into t3 rtotal: Context set the second mode Orientation. Activities for this mode are t2 ke,1, ke,2,ke,3: Explored recall set kf,1, kf,2,kf,3: Fokused recall set Navigation, Focus and Inspection. The user is now in t1 kv,1, kv,2,kv,3: Verified recall set the ability to change the visual as well as the informational focal point in an information visualisation of the dialogue context. The mode Assessment is reached, if the user finds Figure 2: Sequence of Separate Queries objects of interest during his ispection. For this mode the activities Evaluation and Store are available. They help process (see figure 3). In a next step we will evaluate this to express the users appreciation of relevance and to define prototype. the identified recall set. Beside different models for information searching ([1], [13], [5]) it was [15] who combined these approaches in a new model. Based on idea we can enhance this model with our activities and interactive modes (see figure 1). Interactive Orientation Interactive Orientation Session n Navigation Session 1 Navigation Focus Inspection Focus Access Situational Exploration activities Inspection Assessment Evaluation Store Access Situational Exploration activities Assessment Evaluation Store Figure 1: Enhanced model of Spink Figure 3: Screenshot of a prototype 2.2 Visualization The past research ([14], [9]) showed that information visu- 2.3 System support alization is an important concept for the cognitive support In order to support the user during the search task, sys- of the user. [3] said: ”‘Visual interfaces to IR systems ex- tems should be proactive ([10], [16]). To be able to actually ploit powerful human vision and spatial cognition to help hu- support and evaluate our model we need a system which mans mentally organize and electronically access and man- meets the following demands. The system age large, complex information spaces. The aim is to shift • should fundamentally support the interaction model, the user’s mental load from slow reading to faster perceptual processes such as visual pattern recognition.”’. • should map the described activities to support the This statement leads us to the second aspect of our posi- user, tion. If we understand search as a process, whose progres- sion fills our context, then we need also support, in order to • should enable the quantitative and qualitative evalua- understand and interpret this context. So the visualization tion of the model, of results must go beyond the usual measure. Especially the • and should be highly flexible and extensible to inte- different sets of information objects shown in [7] seems to be grate new visualisation technics. useful to visualize (see figure 2). The user needs a portfolio of visualization tools which approach his cognitive abilities. Following the formal description of the information di- Furthermore, the user must be able to get the full control alogue and given the demands we want to introduce the of his search history and the developed information context. Daffodil-system as an experimental system for further de- By logging all activities and the sets of information objects velopment and evaluation of the above described model. It resulting from it, we are able to get a first formal overview provides already, up to a certain extend, the demand for of our context. mapping the user activities to existing available tools. A first prototype is developed which visualize the differ- With the information of our context model including the ent sets of information objects during an information search search path we identifed the following challenges: Relevance Feedback The users implicit and explicit rel- [5] P. Ingwersen. Cognitive perspectives of information evance assessments must be captured and related to retrieval interaction: elements of a cognitive ir theory. possible relevant documents. Journal of Documentation, 52:3–50, 1996. [6] C. C. Kuhlthau. Longitudinal case studies of the Search strategy With the help of the user or by moni- information search process of users in libraries. Library toring the activities the system must provide different & Information Science Research, 10:257–304, 1988. search strategies to raise effciency. [7] P. Landwich, T. Vogel, C.-P. Klas, and M. Hemmje. Collaborative recommendations By logging many dif- Supporting patent retrieval in the context of ferent searches in form of a set of activities, it is possi- innovation-processes by means of information ble to support a user through collaborative recommen- visualisation. In Proceedings of ECKM 2008, 2008. dations. Analyzing a new search from the beginning, [8] P. Landwich, T. Vogel, C.-P. Klas, and M. Hemmje. the system is able to identify similar stored search pro- Model to support patent retrieval in the context of cesses. If this knowledge is visualized for the user, he innovation-processes by means of dialogue and could get benefit for his own search task. information visualisation. Electronic Journal of Knowledge Management, 7:87–98, 1 2009. 3. CONCLUSIONS http://www.ejkm.com/volume-7/v7-1/v7-i1-art9.htm. The idea of this position paper is to support users within [9] L. Nowell, E. Hetzler, and T. Tanasse. Change a search task by logging all activities between the user and blindness in information visualization: A case study. the system. For this, we are able to visualize the context In INFOVIS ’01: Proceedings of the IEEE Symposium and make it cognitive perceptible. Furthemore, we are able on Information Visualization 2001 (INFOVIS’01), to draw conclusions from this activities. This captured in- page 15, Washington, DC, USA, 2001. IEEE formation represents the basis to further understand and Computer Society. support the user. Such support could be done through rec- [10] R. Oppermann. Adaptive user support: ergonomic ommendation via implicit relevance feedback as well as col- design of manually and automatically adaptable laborative recommendations through other users in a similar software. L. Erlbaum Associates Inc., Hillsdale, NJ, situation. We think, that given the context model within USA, 1994. the Daffodil-Framework, we are able to understand and cat- [11] N. Pharo. A new model of information behaviour egories user behavior and provide solid data to support sys- based on the search situation transition schema. Inf. tem oriented IR evaluation, e.g. based on user simulation. Res., 10(1), 2004. We currently investigate and evaluate our research using [12] D. E. Rose. Reconciling information-seeking behavior the Daffodil - framework ([4]) as an experimental system. In with search user interfaces for the web. J. Am. Soc. order to evaluate the listed aspects, we momentarily work Inf. Sci. Technol., 57(6):797–799, 2006. on the following projects: [13] T. Saracevic. The stratified model of information • Task manager: A tool to capture and log all activities retrieval interaction: Extension and applications. In and resulting sets of information objects of a search Proceedings of the American Society for Information task over more then one session. Science, volume 34, pages 313–327, 1997. [14] B. Shneiderman. The eyes have it: A task by data • Visualization: Visualize the context and search path type taxonomy for information visualizations. In VL with help of venn diagrams. ’96: Proceedings of the 1996 IEEE Symposium on • Relevance feedback: Interpretation of activities as im- Visual Languages, pages 336–343, Washington, DC, plicit relevance feedback with term suggestions and re- USA, 1996. IEEE Computer Society. ranked result lists. [15] A. Spink. A user-centered approach to evaluating human interaction with web search engines: an 4. REFERENCES exploratory study. In Information Processing and [1] N. J. Belkin, C. Cool, A. Stein, and U. Thiel. Cases, Management, pages 401–426, 2002. scripts, and information-seeking strategies: On the [16] M. Twidale, D. Nichols, M. B. Twidale, and D. M. design of interactive information retrieval systems. In Nichols. Collaborative browsing and visualisation of Arbeitspapiere der GMD. GMD, Sankt Augustin, the search process. In In Proceedings of ELVIRA-96, November 1994. Milton Keynes, pages 48–7, 1996. [2] R. . B. H. Belkin, N. J.; Oddy. Ask for information [17] Y. Xu. The dynamics of interactive information retrieval. Journal of Documentation, 38:61–71 (Teil 1) retrieval behavior, part i: An activity theory & 145–164 (Teil 2), 1982. perspective. J. Am. Soc. Inf. Sci. Technol., [3] K. Börner and C. Chen. Visual interfaces to digital 58(7):958–970, 2007. libraries: Motivation, utilization, and socio-technical challenges. In Visual Interfaces to Digital Libraries, pages 1–12, London, UK, 2002. Springer-Verlag. [4] N. Fuhr, C.-P. Klas, A. Schaefer, and P. Mutschke. Daffodil: An integrated desktop for supporting high-level search activities in federated digital libraries. In Research and Advanced Technology for Digital Libraries. 6th European Conference, ECDL 2002, pages 597–612, Heidelberg et al., 2002. Springer.