Filters








1,919 Hits in 6.7 sec

Searching, Learning, and Subtopic Ordering: A Simulation-based Analysis [article]

Arthur Câmara, David Maxwell, Claudia Hauff
2022 arXiv   pre-print
one subtopic at a time, or shallowly cover several subtopics?).  ...  Complex search tasks - such as those from the Search as Learning (SAL) domain - often result in users developing an information need composed of several aspects.  ...  (ϕ) strategies for learning-oriented search tasks affect the search behaviour of simulated agents?  ... 
arXiv:2201.11181v1 fatcat:hctzy3kklrel5pgr4mn26lnic4

UFO-BLO: Unbiased First-Order Bilevel Optimization [article]

Valerii Likhosherstov, Xingyou Song, Krzysztof Choromanski, Jared Davis, Adrian Weller
2021 arXiv   pre-print
Bilevel optimization (BLO) is a popular approach with many applications including hyperparameter optimization, neural architecture search, adversarial robustness and model-agnostic meta-learning.  ...  We address this concern by proposing a new FO-BLO-based unbiased estimate of outer-level gradients, enabling us to theoretically guarantee this convergence, with no harm to memory and expected time complexity  ...  search, adversarial robustness and gradient-based meta-learning methods (MAML), which are used in robotics, language, and vision.  ... 
arXiv:2006.03631v2 fatcat:zljisn2jonh2bfz44nknpvflxa

Order Statistics and Item Bank Analysis in Computer Adaptive Testing

J. Sua'rez, A. Franco, R.A. Santos
2013 Procedia Technology - Elsevier  
The experimental results are obtained through a simulation environment that takes into account the definition of the structure of an item bank, the definition of a testing subject and the definition of  ...  An item bank is a deposit of this kind of structures and, in this paper, the item bank structure is defined in terms of statistical indexes arising from the onedimensional Order Statistics Theory, namely  ...  Appendix A introduces in a greater detail some important aspects of Order Statistics Theory and the way this theory relates with the field of item bank analysis.  ... 
doi:10.1016/j.protcy.2013.04.034 fatcat:3lridfee3zdftb2bd43mfqemjm

Debiasing a First-order Heuristic for Approximate Bi-level Optimization [article]

Valerii Likhosherstov, Xingyou Song, Krzysztof Choromanski, Jared Davis, Adrian Weller
2021 arXiv   pre-print
To address this complexity, an earlier first-order method (FOM) was proposed as a heuristic that omits second derivative terms, yielding significant speed gains and requiring only constant memory.  ...  We further demonstrate a rich family of examples where FOM-based SGD does not converge to a stationary point of the ABLO objective.  ...  Provable guar- antees for gradient-based meta-learning.  ... 
arXiv:2106.02487v2 fatcat:fbhrq5xd6fbwnfpcmjr6ddta24

Does the Use of Smart Board Increase Students' Higher Order Thinking Skills (HOTS)?

Abdul Halim Abdullah, Soh Hon Mun, Mahani Mokhtar, Mohd Hilmi Hamzah, Zakiah Mohamad Ashari, Dayana Farzeeha Ali, Norazrena Abu Samah, Nurul Farhana Jumaat, Nor Hasniza Ibrahim, Johari Surif, Sharifah Nurarfah S. Abd Rahman
2020 IEEE Access  
The results of the analysis suggest that there is a statistically significant difference between the ALuSB program, an active learning instruction and a conventional learning method in enhancing each level  ...  INDEX TERMS Active learning, higher order thinking skills, smart board.  ...  By using the Search Tool, the students could also translate a word or a sentence into the target language, look it up in the dictionary, and search images for it.  ... 
doi:10.1109/access.2020.3042832 fatcat:a7mdsnmffnc63pjqyxd365xgbq

GDESA: Greedy Diversity Encoder with Self-Attention for Search Results Diversification

Xubo Qin, Zhicheng Dou, Yutao Zhu, Ji-Rong Wen
2022 ACM Transactions on Information Systems  
Based on a self-attention encoder-decoder structure and an RNN-based document selection component, the model can simultaneously leverage both the global interactions among all the documents and the interactions  ...  Experimental results show that GDESA outperforms previous methods that rely just on global interactions, and our further analysis demonstrates that using both global interactions and document selection  ...  Section 5 contains the description of the sampling and optimization process, the analysis of how self-attention works in search result diversiication task, the analysis of time complexity, and the comparison  ... 
doi:10.1145/3544103 fatcat:ftsy3jv3qnhjjmmga36ffggbvu

The dynamics of interactive information retrieval, Part II: An empirical study from the activity theory perspective

Yunjie Xu, Chengliang Liu
2007 Journal of the American Society for Information Science and Technology  
The authors' experimental simulation of 81 participants in one search session indicates the propositions are largely supported. Their findings indicate IIR behavior is planned.  ...  Activity theory is a powerful theoretical instrument to untangle the "complications." Based on activity theory, a comprehensive framework is proposed in Part I (Y.  ...  Had the subneeds of the search task not required temporal ordering (imagine if the search topic was to study a certain period of French and German history), users might start with any subtopic and not  ... 
doi:10.1002/asi.20574 fatcat:skcjb25lpvejlpbv7xffgeqqtm

Learning to diversify web search results with a Document Repulsion Model

Jingfei Li, Yue Wu, Peng Zhang, Dawei Song, Benyou Wang
2017 Information Sciences  
Although there have been existing learning based diversity search methods, they often involve an iterative sequential selection process in the ranking process, which is computationally complex and time  ...  Further, an efficiency analysis shows that the proposed DRM has a lower computational complexity than the state of the art learning-to-diversify methods.  ...  and sort them in a descending order.  ... 
doi:10.1016/j.ins.2017.05.027 fatcat:g76etg5vm5grfpplfghtzwhj2e

Leveraging Dynamic Query Subtopics for Time-Aware Search Result Diversification [chapter]

Tu Ngoc Nguyen, Nattiya Kanhabua
2014 Lecture Notes in Computer Science  
A key idea is to re-rank search results based on the freshness and popularity of subtopics.  ...  Search result diversification is a common technique for tackling the problem of ambiguous and multi-faceted queries by maximizing query aspects or subtopics in a result list.  ...  The analysis study is based on two data sources, namely, query logs and a temporal document collection, where time information is available.  ... 
doi:10.1007/978-3-319-06028-6_19 fatcat:nvoyghtlenhgbfhqn3kpjdw7qq

Structural Learning of Diverse Ranking [article]

Yadong Zhu, Yanyan Lan, Jiafeng Guo, Xueqi Cheng
2015 arXiv   pre-print
Relevance and diversity are both crucial criteria for an effective search system. In this paper, we propose a unified learning framework for simultaneously optimizing both relevance and diversity.  ...  Specifically, the problem is formalized as a structural learning framework optimizing Diversity-Correlated Evaluation Measures (DCEM), such as ERR-IA, a-NDCG and NRBP.  ...  Although xQuAD list and PM-2 list all utilize the official subtopics as explicit query aspects to simulate their best-case scenarios, their performances are still much lower than learning-based approaches  ... 
arXiv:1504.04596v2 fatcat:n2amn3mcljcnbh2pwce6sre3zm

Understanding the Query: THCIB and THUIS at NTCIR-10 Intent Task

Junjun Wang, Guoyu Tang, Yunqing Xia, Qiang Zhou, Thomas Fang Zheng, Qinan Hu, Sen Na, Yaohai Huang
2013 NTCIR Conference on Evaluation of Information Access Technologies  
3) LDA is applied to discover explicit subtopic candidates within search results. (4) Sense based subtopic clustering and entity analysis are conducted to cluster the subtopic candidates so as to discover  ...  This paper presents the intent mining system developed by THCIB and THUIS, which is capable of understanding English and Chinese query respectively, with four types of context: query, knowledge base, search  ...  We can further discover intents by clustering the subtopic candidates, and design a unified model to rank the intents and subtopics based on both relevance and diversity.  ... 
dblp:conf/ntcir/WangTXZZHNH13 fatcat:d7ybhfhkz5cnhmca4efqjzmquy

Improve the orders picking in e-commerce by using WMS data and BigData analysis
Poboljšanje komisioniranja naloga kod e-trgovine korišćenjem analize WMS i BigData podataka

Augustyn Lorenc, Aurelija Burinskiene
2021 FME Transaction  
The method delivered by authors could be in a typical warehouse, where forklifts and employees do the order picking process.  ...  Such solution could be used for everyday analysis and planning the allocation of products.  ...  Base on the presented formulas, for each order the picking time is calculated.  ... 
doi:10.5937/fme2101233l fatcat:daq5znhuirgsfdq2uxo33p2pbm

Searching to Learn with Instructional Scaffolding [article]

Arthur Câmara, Nirmal Roy, David Maxwell, Claudia Hauff
2021 arXiv   pre-print
queries with relevant subtopics; (ii) CURATED_SC, the presenting of a manually curated static list of relevant subtopics on the search engine result page; and (iii) FEEDBACK_SC, which projects real-time  ...  To investigate the effectiveness of these approaches with respect to human learning, we conduct a user study (N=126) where participants were tasked with searching and learning about topics such as 'genetically  ...  This is a constant that is determined based on the search session length, and the number of subtopics present.  ... 
arXiv:2111.14584v1 fatcat:zg6vnzi4fzf2xpsexya4vvz7kq

Automatic Segmentation of Multiparty Dialogue

Pei-yun Hsueh, Johanna D. Moore, Steve Renals
2006 Conference of the European Chapter of the Association for Computational Linguistics  
Examination of the effect of features shows that predicting top-level and predicting subtopic boundaries are two distinct tasks: (1) for predicting subtopic boundaries, the lexical cohesion-based approach  ...  alone can achieve competitive results, (2) for predicting top-level boundaries, the machine learning approach that combines lexical-cohesion and conversational features performs best, and (3) conversational  ...  Acknowledgements Many thanks to Jean Carletta for her invaluable help in managing the data, and for advice and comments on the work reported in this paper.  ... 
dblp:conf/eacl/HsuehMR06 fatcat:nzjlidze4vgadjp6ksxepjgpfm

RLIRank: Learning to Rank with Reinforcement Learning for Dynamic Search

Jianghong Zhou, Eugene Agichtein
2020 Proceedings of The Web Conference 2020  
To address this problem, we introduce a novel reinforcement learning-based approach, RLIrank.  ...  Then, we implement a new Learning to Rank (LTR) model for each iteration of the dynamic search, using a recurrent Long Short Term Memory neural network (LSTM), which estimates the gain for each next result  ...  Each query has several subtopics, and the documents have a relevant score for some subtopics provided by a Jig user simulator.  ... 
doi:10.1145/3366423.3380047 dblp:conf/www/ZhouA20 fatcat:hivnsj2zlbdkjcbnhttmc3gh4y
« Previous Showing results 1 — 15 out of 1,919 results