Synthesizing Aspect-Driven Recommendation Explanations from Reviews

Proceedings of the Twenty-Ninth International Joint Conference on Artificial Intelligence ◽

10.24963/ijcai.2020/336 ◽

2020 ◽

Author(s):

Trung-Hoang Le ◽

Hady W. Lauw

Keyword(s):

Text Generation ◽

Model Experiments ◽

Product Categories ◽

Review Summarization ◽

Explainable Recommendation

Explanations help users make sense of recommendations, increasing the likelihood of adoption. Existing approaches to explainable recommendations tend to rely on rigidly standardized templates, only allowing fill-in-the-blank aspect-level sentiments. For more flexible, literate, and varied explanations that cover various aspects of interest, we propose to synthesize an explanation by selecting snippets from reviews to optimize representativeness and coherence. To fit the target user's aspect preferences, we contextualize the opinions based on a compatible explainable recommendation model. Experiments on datasets of varying product categories showcase the efficacies of our method as compared to baselines based on templates, review summarization, selection, and text generation.

Download Full-text

DSGPT: Domain-Specific Generative Pre-Training of Transformers for Text Generation in E-commerce Title and Review Summarization

Proceedings of the 44th International ACM SIGIR Conference on Research and Development in Information Retrieval ◽

10.1145/3404835.3463037 ◽

2021 ◽

Author(s):

Xueying Zhang ◽

Yunjiang Jiang ◽

Yue Shang ◽

Zhaomeng Cheng ◽

Chi Zhang ◽

...

Keyword(s):

Text Generation ◽

Domain Specific ◽

Review Summarization

Download Full-text

Cognitive Components of Ingroup Projection

Social Psychology ◽

10.1027/1864-9335/a000156 ◽

2014 ◽

Vol 45 (1) ◽

pp. 15-30 ◽

Cited By ~ 6

Author(s):

Maya Machunsky ◽

Thorsten Meiser

Keyword(s):

Mental Representation ◽

Cognitive Representation ◽

Social Categories ◽

Projection Model ◽

Minimal Group ◽

Model Experiments ◽

Superordinate Category ◽

Cognitive Components ◽

Per Se

This research investigated whether relative ingroup prototypicality (i.e., the tendency to perceive one’s own ingroup as more prototypical of a superordinate category than the outgroup) can result from a prototype-based versus exemplar-based mental representation of social categories, rather than from ingroup membership per se as previously suggested by the ingroup projection model. Experiments 1 and 2 showed that a prototype-based group was perceived as more prototypical of a superordinate category than an exemplar-based group supporting the hypothesis that an intergroup context is not necessary for biased prototypicality judgments. Experiment 3 introduced an intergroup context in a minimal-group-like paradigm. The findings demonstrated that both the kind of cognitive representation and motivational processes contribute to biased prototypicality judgments in intergroup settings.

Download Full-text

Some Contributions to Hydraulic Model Experiments in Coastal Engineering

Coastal Engineering 1966 ◽

10.1061/9780872620087.073 ◽

1967 ◽

Author(s):

Shoitiro Hayami ◽

Tojiro Ishihara ◽

Yuichi Iwagaki

Keyword(s):

Hydraulic Model ◽

Coastal Engineering ◽

Model Experiments

Download Full-text

The Computational Sublime in Nick Montfort's ‘Round’ and ‘All the Names of God’

CounterText ◽

10.3366/count.2015.0027 ◽

2015 ◽

Vol 1 (3) ◽

pp. 348-365 ◽

Cited By ~ 2

Author(s):

Mario Aquilina

Keyword(s):

Computer Code ◽

Computer Programs ◽

Literary Text ◽

Text Generation ◽

Mathematical Concepts ◽

Literary Space

What if the post-literary also meant that which operates in a literary space (almost) devoid of language as we know it: for instance, a space in which language simply frames the literary or poetic rather than ‘containing’ it? What if the countertextual also meant the (en)countering of literary text with non-textual elements, such as mathematical concepts, or with texts that we would not normally think of as literary, such as computer code? This article addresses these issues in relation to Nick Montfort's #!, a 2014 print collection of poems that presents readers with the output of computer programs as well as the programs themselves, which are designed to operate on principles of text generation regulated by specific constraints. More specifically, it focuses on two works in the collection, ‘Round’ and ‘All the Names of God’, which are read in relation to the notions of the ‘computational sublime’ and the ‘event’.

Download Full-text

EMPIRICAL INVESTIGATION ON IMPULSIVE PURCHASE BEHAVIOR: THE INTERPLAY BETWEEN PRODUCT CATEGORIES AND MARKETING ACTIVITIES

Global Fashion Management Conference ◽

10.15444/gmc2018.11.07.04 ◽

2018 ◽

Vol 2018 ◽

pp. 1400-1402

Author(s):

Sungjoon Nam ◽

◽

Minki Kim ◽

Sang-Hoon Kim

Keyword(s):

Empirical Investigation ◽

Purchase Behavior ◽

Product Categories ◽

Marketing Activities

Download Full-text

Influence of Rim Run-Out on the Nonuniformity of Tire-Wheel Assemblies

Tire Science and Technology ◽

10.2346/1.2139538 ◽

1994 ◽

Vol 22 (2) ◽

pp. 99-120 ◽

Cited By ~ 1

Author(s):

T. B. Rhyne ◽

R. Gall ◽

L. Y. Chang

Keyword(s):

Finite Element ◽

Radial Force ◽

Membrane Model ◽

Model Experiments ◽

Force Variation ◽

One To One ◽

The One

Abstract An analytical membrane model is used to study how wheel imperfections are converted into radial force variation of the tire-wheel assembly. This model indicates that the radial run-out of the rim generates run-out of the tire-wheel assembly at slightly less than the one to one ratio that was expected. Lateral run-out of the rim is found to generate radial run-out of the tire-wheel assembly at a ratio that is dependent on the tire design and the wheel width. Finite element studies of a production tire validate and quantify the results of the membrane model. Experiments using a specially constructed precision wheel demonstrate the behavior predicted by the models. Finally, a population of production tires and wheels show that the lateral run-out of the rims contribute a significant portion to the assembly radial force variation. These findings might be used to improve match-mounting results by taking lateral rim run-out into account.

Download Full-text