Medical Semantic Similarity with a Neural Language Model

By Lance De Vine, Guido Zuccon, Bevan Koopman, Laurianne Sitbon, Peter D. Bruza. 2014

161 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Molecular Biology

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

1. Matthew Lease
2. Jimeng Sun
3. Ye Zhang
4. Byron C. Wallace
5. Edward Choi
6. Mohammad Taha Bahadori
7. Guido Zuccon
8. Brandon Dang

In Molecular Biology

1. Edward Choi
2. Jimeng Sun
3. Maarten de Rijke
4. Byron C. Wallace
5. Xu Min
6. Ting Chen
7. Zhaochun Ren

Citations in the corpus

• Patient outcome prediction via convolutional neural networks based on multi-granularity medical concept embedding (2017, methodology): “[19] first extracted UMLS concepts from two sets of free texts, clinical patient records and medical journal abstracts and then learned the embedding using documents obtained by concatenating all of the extracted concepts.”
• Multi-layer Representation Learning for Medical Concepts (2016, background): “…|ci) + ∑ −w≤k≤w,k 6=0 −x>t+k log ŷt − (1− xt+k)> log(1− ŷt) } By combining the two objective functions we learn both code representations and visit representations from the same source of patient visit records, exploiting both intra-visit co-occurrence information as well as inter-visit…”
• Neural Attentive Session-based Recommendation (2017, methodology): “Deep learning has recently been applied very successfully in areas such as image recognition [8, 17], speech recognition [2, 7, 13] and neural language processing [5, 18, 30, 37, 38].”
• Neural information retrieval: at the end of the early years (2017, methodology): “Experiments conducted on ClueWeb09-Cat-B with TREC 2010–2012 Web Track topics show QEM outperforms Gao et al. (2010)’s concept translation model (CTM) (statistically significant differences), and SSI (occasionally statistically significant differences).”
• Integrating and Evaluating Neural Word Embeddings in Information Retrieval (2015, methodology): “Typical values used in computational linguistics tasks are 100-600 for the latent dimensionality of the word embeddings and 5-10 for the context window [21, 23, 12, 25].”