Learning Text Similarity with Siamese Recurrent Networks
By Paul Neculoiu, Maarten Versteegh, Mihai Rotaru. Rep4NLP@ACL, 2016
334 citations recorded by the source corpus.
This paper presents a deep architecture for learning a similarity metric on variable-length character sequences. The model combines a stack of character-level bidirectional LSTM’s with a Siamese architecture. It learns to project variable-length strings into a fixed-dimensional embedding space by using only information about the similarity between pairs of strings. This model is applied to the task of job title normalization based on a manually annotated taxonomy. A small data set is incrementally expanded and augmented with new sources of variance. The model learns a representation that is selective to differences in the input that reflect semantic differences (e.g., “Java developer” vs. “HR manager”) but also invariant to non-semantic string differences (e.g., “Java de-veloper” vs. “Java programmer”).
Subfields: Artificial Intelligence
View Paper: https://www.semanticscholar.org/paper/35b11ac652646c70a559f7ae29295e1d5de09a80
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Artificial Intelligence
Citations in the corpus
- NERO: A Neural Rule Grounding Framework for Label-Efficient Relation Extraction (2020, methodology): “We use the contrastive loss [22] for this objective.”
- Neural Snowball for Few-Shot Relation Learning (2019, background): “They are exploited for one/few-shot learning (Koch et al., 2015; Yuan et al., 2017; Sung et al., 2018), object tracking (Bertinetto et al., 2016) and measuring text similarities (Neculoiu et al., 2016; Mueller and Thyagarajan, 2016).”
- Learning from Explanations with Neural Execution Tree (2019, positive): “To solve this problem, we borrow the idea of word retrofitting (Faruqui et al., 2014) and adopt a contrastive loss (Neculoiu et al., 2016) to incorporate semantic knowledge in training.”
- Learning to Annotate: Modularizing Data Augmentation for Text Classifiers with Natural Language Explanations (2019, methodology): “To solve this problem, we borrow the idea of word retrofitting (Faruqui et al., 2014) and adopt a contrastive loss (Neculoiu et al., 2016) to incorporate semantic knowledge in training.”
- Deep Learning--based Text Classification (2020, methodology): “[120] present a similar model that uses character-level Bi-LSTMs for f1 and f2, and the”
- Semantic Text Matching for Long-Form Documents (2019, methodology): “For instance, some studies [2, 27, 32, 33, 45, 46] exploit RNNs to model the sequential information of documents while other work [19, 38, 46, 54] applies CNNs to capture the important local messages in documents for semantic text matching.”
- Adversarial Domain Adaptation for Variational Neural Language Generation in Dialogue Systems (2018, methodology): “We further utilize Siamese recurrent architectures (Neculoiu et al., 2016) for learning sentence similarity, in which the architecture allows us to learn useful representations with limited supervision.”
- Sampling-bias-corrected neural modeling for large corpus item recommendations (2019, background): “Building neural networks with two towers has recently become a popular approach in several natural language tasks including modeling sentence similarities [31], response suggestions [24], and text-based information retrieval [17, 37].”
- Deep Learning for Entity Matching: A Design Space Exploration (2018, methodology): “Table 13: Comparing the F1 accuracy of the Hybrid and RNN models in our submission with the model proposed in [56] for an NLP task.”
- CAIL2019-SCM: A Dataset of Similar Case Matching in Legal Domain (2019, background): “The Siamese structure (Bromley et al., 1994) for metric learning achieve great success and is widely applied (Amiri et al., 2016; Liu et al., 2018; Mueller and Thyagarajan, 2016; Neculoiu et al., 2016; Wan et al., 2016; He et al., 2015).”