61 |
SEAGLE: A platform for comparative evaluation of semantic encoders for information retrieval
|
|
|
|
BASE
|
|
Show details
|
|
63 |
Specializing distributional vectors of all words for lexical entailment
|
|
|
|
BASE
|
|
Show details
|
|
64 |
How to (properly) evaluate cross-lingual word embeddings: On strong baselines, comparative analyses, and some misconceptions
|
|
|
|
BASE
|
|
Show details
|
|
65 |
Cross-lingual semantic specialization via lexical relation induction
|
|
|
|
BASE
|
|
Show details
|
|
66 |
Generalized tuning of distributional word vectors for monolingual and cross-lingual lexical entailment
|
|
|
|
BASE
|
|
Show details
|
|
67 |
SenZi: A sentiment analysis lexicon for the latinised Arabic (Arabizi)
|
|
|
|
BASE
|
|
Show details
|
|
68 |
Informing unsupervised pretraining with external linguistic knowledge
|
|
|
|
BASE
|
|
Show details
|
|
69 |
Do we really need fully unsupervised cross-lingual embeddings?
|
|
|
|
BASE
|
|
Show details
|
|
70 |
Are we consistently biased? Multidimensional analysis of biases in distributional word vectors
|
|
|
|
BASE
|
|
Show details
|
|
71 |
Unsupervised Cross-Lingual Information Retrieval using Monolingual Data Only ...
|
|
|
|
BASE
|
|
Show details
|
|
72 |
Unsupervised Cross-Lingual Information Retrieval Using Monolingual Data Only ...
|
|
|
|
BASE
|
|
Show details
|
|
73 |
Adversarial Propagation and Zero-Shot Cross-Lingual Transfer of Word Vector Specialization ...
|
|
|
|
Abstract:
Semantic specialization is the process of fine-tuning pre-trained distributional word vectors using external lexical knowledge (e.g., WordNet) to accentuate a particular semantic relation in the specialized vector space. While post-processing specialization methods are applicable to arbitrary distributional vectors, they are limited to updating only the vectors of words occurring in external lexicons (i.e., seen words), leaving the vectors of all other words unchanged. We propose a novel approach to specializing the full distributional vocabulary. Our adversarial post-specialization method propagates the external lexical knowledge to the full distributional space. We exploit words seen in the resources as training examples for learning a global specialization function. This function is learned by combining a standard L2-distance loss with an adversarial loss: the adversarial component produces more realistic output vectors. We show the effectiveness and robustness of the proposed method across three ... : Accepted at EMNLP 2018 ...
|
|
Keyword:
Computation and Language cs.CL; FOS Computer and information sciences
|
|
URL: https://dx.doi.org/10.48550/arxiv.1809.04163 https://arxiv.org/abs/1809.04163
|
|
BASE
|
|
Hide details
|
|
74 |
Post-Specialisation: Retrofitting Vectors of Words Unseen in Lexical Resources ...
|
|
|
|
BASE
|
|
Show details
|
|
75 |
A Resource-Light Method for Cross-Lingual Semantic Textual Similarity ...
|
|
|
|
BASE
|
|
Show details
|
|
76 |
Post-Specialisation: Retrofitting Vectors of Words Unseen in Lexical Resources ...
|
|
|
|
BASE
|
|
Show details
|
|
77 |
Unsupervised Cross-Lingual Information Retrieval Using Monolingual Data Only
|
|
|
|
BASE
|
|
Show details
|
|
78 |
ArguminSci: a tool for analyzing argumentation and rhetorical aspects in scientific writing
|
|
|
|
BASE
|
|
Show details
|
|
80 |
Investigating the role of argumentation in the rhetorical analysis of scientific publications with neural multi-task learning models
|
|
|
|
BASE
|
|
Show details
|
|
|
|