Tagging Address Queries in Maps Search

Mokhtari, Shekoofeh; Mahmoody, Ahmad; Yankov, Dragomir; Xie, Ning

doi:10.1609/aaai.v33i01.33019547

Cited by 10 publications

(9 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In recent years, new methods [2], [3] utilizing the power of neural networks have been proposed as solutions for the address parsing problem. Using a single hidden layer feedforward model, [6] achieved good performance.…”

Section: Related Workmentioning

confidence: 99%

“…In addition, the model's predictions are put through a rulebased validation step to make sure that they fit known patterns. In contrast, [3] proposed a deep learning approach based on the use of RNN and did not use any pre or post-processing. Their experiments focused on comparing the performance of both unidirectional and bidirectional vanilla RNN and Long-Short Term Memory Models (LSTM) [9], as well as a Seq2Seq.…”

Section: Related Workmentioning

confidence: 99%

“…This library uses a CRF-based model trained with an averaged Perceptron for scalability. The model was trained on data from each country in the world and was able to achieve a 99.45 % full parse accuracy 3 , thus defining a new state-of-theart 4 . However, this requires putting addresses through a heavy pre-processing pipeline before feeding them to the prediction model.…”

Section: Related Workmentioning

confidence: 99%

“…https://github.com/openvenues/libpostal3 The accuracy was computed considering the entire sequence and was not focused on individual tokens 4. For a comparison of some of our models with Libpostal, visit our previous article[7].…”

mentioning

confidence: 99%

See 3 more Smart Citations

Multinational Address Parsing: A Zero-Shot Evaluation

Yassine¹,

Beauchemin²,

Laviolette³

et al. 2021

Preprint

View full text Add to dashboard Cite

Address parsing consists of identifying the segments that make up an address, such as a street name or a postal code. Because of its importance for tasks like record linkage, address parsing has been approached with many techniques, the latest relying on neural networks. While these models yield notable results, previous work on neural networks has only focused on parsing addresses from a single source country. This paper explores the possibility of transferring the address parsing knowledge acquired by training deep learning models on some countries' addresses to others with no further training in a zero-shot transfer learning setting. We also experiment using an attention mechanism and a domain adversarial training algorithm in the same zero-shot transfer setting to improve performance. Both methods yield state-of-the-art performance for most of the tested countries while giving good results to the remaining countries. We also explore the effect of incomplete addresses on our best model, and we evaluate the impact of using incomplete addresses during training. In addition, we propose an open-source Python implementation of some of our trained models 1 .

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

mentioning

confidence: 99%

See 2 more Smart Citations

Multinational Address Parsing: A Zero-Shot Evaluation

Yassine¹,

Beauchemin²,

Laviolette³

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

“…These previous pieces of work did a remarkable job at finding solutions for the challenges linked to the address parsing task. However, most of these approaches either do not take into account parsing addresses from different countries or do so but at the cost of a considerable amount of meta-data and a substantial data pre-processing pipelines [3]- [6].…”

Section: Introductionmentioning

confidence: 99%

Leveraging Subword Embeddings for Multinational Address Parsing

Yassine

Beauchemin

Laviolette

et al. 2020

2020 6th IEEE Congress on Information Science and Technology (CiSt)

View full text Add to dashboard Cite

Address parsing consists of identifying the segments that make up an address such as a street name or a postal code. Because of its importance for tasks like record linkage, address parsing has been approached with many techniques. Neural network methods defined a new state-of-the-art for address parsing. While this approach yielded notable results, previous work has only focused on applying neural networks to achieve address parsing of addresses from one source country. We propose an approach in which we employ subword embeddings and a Recurrent Neural Network architecture to build a single model capable of learning to parse addresses from multiple countries at the same time while taking into account the difference in languages and address formatting systems. We achieved accuracies around 99 % on the countries used for training with no pre-processing nor post-processing needed. In addition, we explore the possibility of transferring the address parsing knowledge attained by training on some countries' addresses to others with no further training. This setting is also called zero-shot transfer learning. We achieve good results for 80 % of the countries (34 out of 41), almost 50 % of which (19 out of 41) is near state-of-the-art performance.

show abstract