Recommender systems are embracing conversational technologies to obtain user preferences dynamically, and to overcome inherent limitations of their static models. A successful Conversational Recommender System (CRS) requires proper handling of interactions between conversation and recommendation. We argue that three fundamental problems need to be solved: 1) what questions to ask regarding item attributes, 2) when to recommend items, and 3) how to adapt to the users' online feedback. To the best of our knowledge, there lacks a unified framework that addresses these problems.In this work, we fill this missing interaction framework gap by proposing a new CRS framework named Estimation-Action-Reflection, or EAR, which consists of three stages to better converse with users. (1) Estimation, which builds predictive models to estimate user preference on both items and item attributes; (2) Action, which learns a dialogue policy to determine whether to ask attributes or recommend items, based on Estimation stage and conversation history; and (3) Reflection, which updates the recommender model when a user rejects the recommendations made by the Action stage. We present two conversation scenarios on binary and enumerated questions, and conduct extensive experiments on two datasets from Yelp and LastFM, for each scenario, respectively. Our experiments demonstrate significant improvements over the state-of-the-art method CRM [32], corresponding to fewer conversation turns and a higher level of recommendation hits. CCS CONCEPTS• Information systems → Users and interactive retrieval; Recommender systems; Personalization; • Human-centered computing → Interactive systems and tools.
Wire arc additive manufacturing (WAAM) offers a potential approach to fabricate large-scale magnesium alloy components with low cost and high efficiency, although this topic is yet to be reported in literature. In this study, WAAM is preliminarily applied to fabricate AZ31 magnesium. Fully dense AZ31 magnesium alloy components are successfully obtained. Meanwhile, to refine grains and obtain good mechanical properties, the effects of pulse frequency (1, 2, 5, 10, 100, and 500 Hz) on the macrostructure, microstructure and tensile properties are investigated. The results indicate that pulse frequency can result in the change of weld pool oscillations and cooling rate. This further leads to the change of the grain size, grain shape, as well as the tensile properties. Meanwhile, due to the resonance of the weld pool at 5 Hz and 10 Hz, the samples have poor geometry accuracy but contain finer equiaxed grains (21 μm) and exhibit higher ultimate tensile strength (260 MPa) and yield strength (102 MPa), which are similar to those of the forged AZ31 alloy. Moreover, the elongation of all samples is above 23%.
Multi-armed bandit algorithms have become a reference solution for handling the explore/exploit dilemma in recommender systems, and many other important real-world problems, such as display advertisement. However, such algorithms usually assume a stationary reward distribution, which hardly holds in practice as users' preferences are dynamic. This inevitably costs a recommender system consistent suboptimal performance. In this paper, we consider the situation where the underlying distribution of reward remains unchanged over (possibly short) epochs and shifts at unknown time instants. In accordance, we propose a contextual bandit algorithm that detects possible changes of environment based on its reward estimation confidence and updates its arm selection strategy respectively. Rigorous upper regret bound analysis of the proposed algorithm demonstrates its learning effectiveness in such a non-trivial environment. Extensive empirical evaluations on both synthetic and real-world datasets for recommendation confirm its practical utility in a changing environment.
Contextual bandit algorithms provide principled online learning solutions to find optimal trade-offs between exploration and exploitation with companion side-information. Most contextual bandit algorithms simply assume the learner would have access to the entire set of features, which govern the generation of payoffs from a user to an item. However, in practice it is challenging to exhaust all relevant features ahead of time, and oftentimes due to privacy or sampling constraints many factors are unobservable to the algorithm. Failing to model such hidden factors leads a system to make constantly suboptimal predictions. In this paper, we propose to learn the hidden features for contextual bandit algorithms. Hidden features are explicitly introduced in our reward generation assumption, in addition to the observable contextual features. A scalable bandit algorithm is achieved via coordinate descent, in which closed form solutions exist at each iteration for both hidden features and bandit parameters. Most importantly, we rigorously prove that the developed contextual bandit algorithm achieves a sublinear upper regret bound with high probability, and a linear regret is inevitable if one fails to model such hidden features. Extensive experimentation on both simulations and large-scale real-world datasets verified the advantages of the proposed algorithm compared with several state-of-the-art contextual bandit algorithms and existing ad-hoc combinations between bandit algorithms and matrix factorization methods.
To increase building rate and save cost, the selective laser melting (SLM) of Ti6Al4V with a high layer thickness (200 μm) and low cost coarse powders (53 μm–106 μm) at a laser power of 400 W is investigated in this preliminary study. A relatively large laser beam with a diameter of 200 μm is utilized to produce a stable melt pool at high layer thickness, and the appropriate scanning track, which has a smooth surface with a shallow contact angle, can be obtained at the scanning speeds from 40 mm/s to 80 mm/s. By adjusting the hatch spacings, the density of multi-layer samples can be up to 99.99%, which is much higher than that achieved in previous studies about high layer thickness selective laser melting. Meanwhile, the building rate can be up to 7.2 mm3/s, which is about 2 times–9 times that of the commercial equipment. Besides, two kinds of defects are observed: the large un-melted defects and the small spherical micropores. The formation of the un-melted defects is mainly attributed to the inappropriate overlap rates and the unstable scanning tracks, which can be eliminated by adjusting the processing parameters. Nevertheless, the micropores cannot be completely eliminated. It is worth noting that the high layer thickness plays a key role on surface roughness rather than tensile properties during the SLM process. Although a sample with a relatively coarse surface is generated, the average values of yield strength, ultimate tensile strength, and elongation are 1050 MPa, 1140 MPa, and 7.03%, respectively, which are not obviously different than those with the thin layer thickness used in previous research; this is due to the similar metallurgical bonding and microstructure.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.