Experimental syntax and the variation of island effects in English and Italian AcknowledgmentsThis work was supported in part by National Science Foundation grants BCS-0843896 and BSC-1347115 to JS. We would like to thank Michela Marchesi for assistance collecting data for the Italian WH-dependencies experiment. We would like to thank Jeremy Hartman, Norbert Hornstein, Luigi Rizzi, and two anonymous reviewers for helpful comments and suggestions on earlier versions of this article. We would also like to thank audiences at Harvard University, Johns Hopkins University, and the University of Illinois Chicago for helpful comments at various stages of the development of this study. All errors remain our own.1 Experimental syntax and the variation of island effects in English and Italian AbstractThe goal of this article is to explore the utility of experimental syntax techniques in the investigation of syntactic variation. To that end, we applied the factorial definition of island effects made available by experimental syntax (e.g., ) to four island types (wh/whether, complex NP, subject, and adjunct), two dependency types (wh-interrogative clause dependencies and relative clause dependencies) and two languages (English and Italian). The results of 8 primary experiments suggest that there is indeed variation across dependency types, suggesting that wh-interrogative clause dependencies and relative clause dependencies cannot be identical at every level of analysis; however, the pattern of variation observed in these experiments is not exactly the pattern of variation previously reported in the literature (e.g., Rizzi 1982). We review six major syntactic approaches to the analysis island effects (Subjacency, CED, Barriers, Relativized Minimality, Structure-building, and Phases) and discuss the implications of these results for these analyses. We also present 4 supplemental experiments testing complex wh-phrases (also called D-linked or lexically restricted wh-phrases) for all four island types using the factorial design in order to tease apart the contribution of dependency type from featural specification. The results of the supplemental experiments confirm that dependency type is the major source of variation, not featural specification, while providing a concrete quantification of what exactly the effect of complex wh-phrases on island effects is.

No abstract

We address the problem of user intent prediction from clickstream data of an e-commerce website via two conceptually different approaches: a hand-crafted feature-based classification and a deep learning-based classification. In both approaches, we deliberately coarse-grain a new clickstream proprietary dataset to produce symbolic trajectories with minimal information. Then, we tackle the problem of trajectory classification of arbitrary length and ultimately, early prediction of limited-length trajectories, both for balanced and unbalanced datasets. Our analysis shows that k-gram statistics with visibility graph motifs produce fast and accurate classifications, highlighting that purchase prediction is reliable even for extremely short observation windows. In the deep learning case, we benchmarked previous state-of-the-art (SOTA) models on the new dataset, and improved classification accuracy over SOTA performances with our proposed LSTM architecture. We conclude with an in-depth error analysis and a careful evaluation of the pros and cons of the two approaches when applied to realistic industry use cases.

The chapter focuses on V3 patterns in West Flemish in which a subject-initial non-inverted V2 clause is preceded by an adverbial adjunct which modifies temporal or modal coordinates of the associated clause, in apparent violation of the V2 constraint. The pattern is not available in many other varieties of Dutch, including Standard Dutch. The chapter summarizes the main distributional and interpretive properties of the initial constituent, focusing on, among other things, the fact that for its interpretation, the initial adjunct cannot be reconstructed to a clause-internal position. On the basis of the distributional and interpretive properties of the initial constituent, it is argued that these V3 patterns are in line with V2 syntax because the initial constituent is extrasentential. The chapter develops the discourse syntax for main clause external constituents and argues that the micro-variation observed can be captured by the hypothesis that there is micro-variation between Standard Dutch and West Flemish in terms of the derivation of subject-initial V2 root clauses.

We address the problem of personalizing query completion in a digital commerce setting, in which the bounce rate is typically high and recurring users are rare. We focus on in-session personalization and improve a standard noisy channel model by injecting dense vectors computed from product images at query time. We argue that image-based personalization displays several advantages over alternative proposals (from data availability to business scalability), and provide quantitative evidence and qualitative support on the effectiveness of the proposed methods. Finally, we show how a shared vector space between similar shops can be used to improve the experience of users browsing across sites, opening up the possibility of applying zero-shot unsupervised personalization to increase conversions. This will prove to be particularly relevant to retail groups that manage multiple brands and/or websites and to multi-tenant SaaS providers that serve multiple clients in the same space.