User modeling

About: User modeling is a research topic. Over the lifetime, 10701 publications have been published within this topic receiving 278012 citations.

User interface construction

Abstract: A system and method for constructing a graphical user interface for an application being accessed by a user are provided. The method includes measuring the user's current work rate, deriving a threshold from the user's current work rate, determining the user's current activity within the application, assigning a value to the user's current activity, accessing a model for the application, the model defining links between activities within the application, assigning values to activities linked within the accessed model to the user's current activity, and displaying a control element in a graphical user interface for each activity with a value above the derived threshold.

Disentangling User Interest and Conformity for Recommendation with Causal Embedding

Abstract: Recommendation models are usually trained on observational interaction data. However, observational interaction data could result from users’ conformity towards popular items, which entangles users’ real interest. Existing methods tracks this problem as eliminating popularity bias, e.g., by re-weighting training samples or leveraging a small fraction of unbiased data. However, the variety of user conformity is ignored by these approaches, and different causes of an interaction are bundled together as unified representations, hence robustness and interpretability are not guaranteed when underlying causes are changing. In this paper, we present DICE, a general framework that learns representations where interest and conformity are structurally disentangled, and various backbone recommendation models could be smoothly integrated. We assign users and items with separate embeddings for interest and conformity, and make each embedding capture only one cause by training with cause-specific data which is obtained according to the colliding effect of causal inference. Our proposed methodology outperforms state-of-the-art baselines with remarkable improvements on two real-world datasets on top of various backbone models. We further demonstrate that the learned embeddings successfully capture the desired causes, and show that DICE guarantees the robustness and interpretability of recommendation.

System and method for tracking web interactions with real time analytics

Abstract: A device, system and method is provided for monitoring a user's interactions with Internet-based programs or documents. Content may be extracted from Internet server traffic according to predefined rules. Extracted content may be associated with a user's Internet interaction. The user's Internet interaction may be stored and indexed. The user's Internet interaction may be analyzed to generate a recommendation provided to a contact center agent while the contact center agent is communicating with said user for guiding the user's interaction, for example, in real-time. Traffic other than Internet server traffic may also be used.

PATS: Realization and user evaluation of an automatic playlist generator.

Abstract: An automatic music playlist generator called PATS (Personalized Automatic Track Selection) creates playlists that aim at suiting a particular listening situation. It uses dynamic clustering in which songs are grouped based on a weighted attribute-value similarity measure. An inductive learning algorithm is used to reveal the weights for attribute-values using user preference feedback. In a controlled user experiment, the quality of PATS-generated and randomly assembled playlists for jazz music was assessed in two listening situations. The two listening situations were “listening to soft music” and “listening to lively music.” Playlist quality was measured by precision (songs that suit the listening situation), coverage (songs that suit the listening situation but that were not already contained in previous playlists) and a rating score. Results showed that PATS playlists contained increasingly more preferred music (increasingly higher precision), covered more preferred music in the collection (higher coverage), and were rated higher than randomly assembled playlists.

Modeling user search behavior for masquerade detection

Abstract: Masquerade attacks are a common security problem that is a consequence of identity theft. This paper extends prior work by modeling user search behavior to detect deviations indicating a masquerade attack. We hypothesize that each individual user knows their own file system well enough to search in a limited, targeted and unique fashion in order to find information germane to their current task. Masqueraders, on the other hand, will likely not know the file system and layout of another user's desktop, and would likely search more extensively and broadly in a manner that is different than the victim user being impersonated. We identify actions linked to search and information access activities, and use them to build user models. The experimental results show that modeling search behavior reliably detects all masqueraders with a very low false positive rate of 1.1%, far better than prior published results. The limited set of features used for search behavior modeling also results in large performance gains over the same modeling techniques that use larger sets of features.