Information Retrieval: Human-Computer Interfaces and Information Access Process

Introduction Information seeking is an imprecise process. –Users have fuzzy understanding of how to achieve their goals –Users select among sources, formulate their queries, review search results, … IR interfaces need to support this process Differences in IR retrieval quality can be dwarfed by differences in interface quality

An Information Task User explicitly or implicitly expresses information need to system System locates information based on expression of need System presents information to user User interacts with information to solve task

Human-Computer Interaction What makes an effective human-computer interface? “Well designed, effective computer systems generate positive feelings of success, competence, mastery, and clarity in the user community. When an interactive system is well-designed, the interface almost disappears, enabling users to concentrate on their work, exploration, or pleasure.” (Ben Shneiderman)

One Set of Design Principles HCI includes a number of lists of design principles (rules of thumb for good design) Selecting list and principles is personal choice and context dependent. Shneiderman’s set includes: –feedback –reversal of actions –Internal locus of control –reduce working memory load –alternative interfaces for novice and expert users

Offer Informative Feedback Comes from early HCI models viewing user interaction as a two-way communication between user and system. Information retrieval includes feedback on –relationship between query and documents retrieved, –relationships among retrieved documents, and –relationships between retrieved documents and metadata.

Reduce Working Memory Load Don’t make users remember – they have enough to do already. Information retrieval interfaces might: –Maintain a history of searches/interactions –Provide browsable information (collection list, index terms, thesaurus, …) –Make suggestions based on user history, knowledge of task, knowledge of collection and retrieved documents.

Alternative Interfaces for Novice and Expert Users The first time user, the casual user, and the regular user need different support. Inverse relationship between ease-of-use and user’s control (e.g. specification of details of ranking approach) Scaffolding used to support users moving between interfaces (the interfaces do not need to be independent)

Role of Visualization Humans are highly attuned to images and visual information. Data (scientific) visualization tends to use physical intuitions of users. Information visualization of abstract concepts (e.g. document space) is more complex.

Information Visualization Techniques Icons and color highlighting Brushing and linking Panning and zooming Focus-plus-context Magic lenses Animation (trees and hierarchies)

Evaluating Interactive Systems Empirical data involving human users is time consuming to gather and difficult to draw universal conclusions from. Evaluation metrics for user interfaces –Time required to learn the system –Time to achieve goals on benchmark tasks –Error rates –Retention of the use of the interface over time –User satisfaction

Models of Interaction Traditional Model: What is wrong? Information need Query Send to System Receive Results Evaluate Results Reformulate Done? Stop No? Yes

Traditional Model of Interaction This model assumes that –the user´s information need is static and –the information seeking process is one of successively refining a query the –ends when until the query retrieves all and only those documents relevant to the original information need.

Traditional Model of Interaction In reality: –Users learn during search process. –Scanning/reading is interlaced with search. –Users browse/examine collection overviews. –User navigation makes near misses valuable.

More Realistic Model Berry-Picking Model –Users’ information needs change as they read/learn. –As a result their queries change. –User needs are not satisfied by a single final set of documents –They are satisfied by selections of information found during the process. The goal is not to generate a query that returns the right information, but for the user to see the right information.

Interfaces Support this Process Given the berry-picking model, the interface should: –allow users to reassess their goals and adjust their search strategy accordingly. –support search strategies by making it easy to follow trails with unanticipated results. –support methods for monitoring the status of the current strategy in relation to the user’s current task and high-level goals. –Some techniques: scanning, querying, navigating, browsing...

Non-Search Parts of Process Information seeking is only one part of users’ full information work process. Users also read, annotate, and analyze documents. One study found 80% of work to be: –Finding trends, –Making comparisons, –Aggregating information, –Identifying a critical subset, –Assessing, and –Interpreting.

Interfaces for More than Search It is convenient to divide the access process into two main components: –search/retrieval –analysis/synthesis Interfaces should allow tight integration of these two activities.

Early IR Interface Studies Modern systems versus old systems. –Full text versus bibliographic citations. –Statistical rankings versus Boolean systems. –End users versus professional intermediaries. –Online collections versus separated collections. –Graphic displays versus text displays. Despite these differences, some general information seeking strategies have been identified that seem to transfer across systems. Some old problems were resolved but new problems were introduced.