2 |
Character Alignment in Morphologically Complex Translation Sets for Related Languages ...
|
|
|
|
BASE
|
|
Show details
|
|
3 |
Composing Byte-Pair Encodings for Morphological Sequence Classification ...
|
|
|
|
BASE
|
|
Show details
|
|
4 |
Variation in Universal Dependencies annotation: A token based typological case study on adpossessive constructions ...
|
|
|
|
BASE
|
|
Show details
|
|
5 |
Corpus evidence for word order freezing in Russian and German ...
|
|
|
|
BASE
|
|
Show details
|
|
7 |
Noise Isn't Always Negative: Countering Exposure Bias in Sequence-to-Sequence Inflection Models ...
|
|
|
|
BASE
|
|
Show details
|
|
8 |
Exhaustive Entity Recognition for Coptic - Challenges and Solutions ...
|
|
|
|
BASE
|
|
Show details
|
|
9 |
Imagining Grounded Conceptual Representations from Perceptual Information in Situated Guessing Games ...
|
|
|
|
BASE
|
|
Show details
|
|
10 |
Attentively Embracing Noise for Robust Latent Representation in BERT ...
|
|
|
|
BASE
|
|
Show details
|
|
13 |
Classifier Probes May Just Learn from Linear Context Features ...
|
|
|
|
BASE
|
|
Show details
|
|
14 |
Seeing the world through text: Evaluating image descriptions for commonsense reasoning in machine reading comprehension ...
|
|
|
|
BASE
|
|
Show details
|
|
16 |
Manifold Learning-based Word Representation Refinement Incorporating Global and Local Information ...
|
|
|
|
BASE
|
|
Show details
|
|
17 |
HMSid and HMSid2 at PARSEME Shared Task 2020: Computational Corpus Linguistics and unseen-in-training MWEs ...
|
|
|
|
BASE
|
|
Show details
|
|
18 |
Multi-dialect Arabic BERT for Country-level Dialect Identification ...
|
|
|
|
BASE
|
|
Show details
|
|
19 |
Autoencoding Improves Pre-trained Word Embeddings ...
|
|
|
|
Abstract:
Prior work investigating the geometry of pre-trained word embeddings have shown that word embeddings to be distributed in a narrow cone and by centering and projecting using principal component vectors one can increase the accuracy of a given set of pre-trained word embeddings. However, theoretically, this post-processing step is equivalent to applying a linear autoencoder to minimise the squared l2 reconstruction error. This result contradicts prior work (Mu and Viswanath, 2018) that proposed to remove the top principal components from pre-trained embeddings. We experimentally verify our theoretical claims and show that retaining the top principal components is indeed useful for improving pre-trained word embeddings, without requiring access to additional linguistic resources or labelled data. ...
|
|
Keyword:
Computer and Information Science; Natural Language Processing; Neural Network
|
|
URL: https://underline.io/lecture/6158-autoencoding-improves-pre-trained-word-embeddings https://dx.doi.org/10.48448/x54c-4398
|
|
BASE
|
|
Hide details
|
|
20 |
Exploring End-to-End Differentiable Natural Logic Modeling ...
|
|
|
|
BASE
|
|
Show details
|
|
|
|