1 |
Negative language transfer in learner English: A new dataset ...
|
|
|
|
BASE
|
|
Show details
|
|
3 |
Parallel sentences mining with transfer learning in an unsupervised setting ...
|
|
|
|
BASE
|
|
Show details
|
|
4 |
Source and Target Bidirectional Knowledge Distillation for End-to-end Speech Translation ...
|
|
|
|
BASE
|
|
Show details
|
|
5 |
Detoxifying Language Models Risks Marginalizing Minority Voices ...
|
|
|
|
BASE
|
|
Show details
|
|
6 |
Domain Adaptation for Arabic Cross-Domain and Cross-Dialect Sentiment Analysis from Contextualized Word Embedding ...
|
|
|
|
BASE
|
|
Show details
|
|
7 |
Knowledge Enhanced Masked Language Model for Stance Detection ...
|
|
|
|
BASE
|
|
Show details
|
|
8 |
LIVE SESSION: 15D-Oral: Phonology, Morphology and Word Segmentation ...
|
|
|
|
BASE
|
|
Show details
|
|
10 |
How low is too low? A monolingual take on lemmatisation in Indian languages ...
|
|
|
|
BASE
|
|
Show details
|
|
11 |
Frustratingly Easy Edit-based Linguistic Steganography with a Masked Language Model ...
|
|
|
|
BASE
|
|
Show details
|
|
12 |
MelBERT: Metaphor Detection via Contextualized Late Interaction using Metaphorical Identification Theories ...
|
|
|
|
BASE
|
|
Show details
|
|
13 |
A Global Past-Future Early Exit Method for Accelerating Inference of Pre-trained Language Models ...
|
|
|
|
Abstract:
Read the paper on the folowing link: https://www.aclweb.org/anthology/2021.naacl-main.162/ Abstract: Early exit mechanism aims to accelerate the inference speed of large-scale pre-trained language models. The essential idea is to exit early without passing through all the inference layers at the inference stage. To make accurate predictions for downstream tasks, the hierarchical linguistic information embedded in all layers should be jointly considered. However, much of the research up to now has been limited to use local representations of the exit layer. Such treatment inevitably loses information of the unused past layers as well as the high-level features embedded in future layers, leading to sub-optimal performance. To address this issue, we propose a novel Past-Future method to make comprehensive pre-dictions from a global perspective. We first take into consideration all the linguistic information embedded in the past layers and further engage the future information which is originally inaccessible ...
|
|
URL: https://underline.io/lecture/19973-a-global-past-future-early-exit-method-for-accelerating-inference-of-pre-trained-language-models https://dx.doi.org/10.48448/s4pe-x386
|
|
BASE
|
|
Hide details
|
|
14 |
DirectProbe: Studying Representations without Classifiers ...
|
|
|
|
BASE
|
|
Show details
|
|
15 |
Challenging distributional models with a conceptual network of philosophical terms ...
|
|
|
|
BASE
|
|
Show details
|
|
17 |
ERNIE-Gram: Pre-Training with Explicitly N-Gram Masked Language Modeling for Natural Language Understanding ...
|
|
|
|
BASE
|
|
Show details
|
|
18 |
Proteno: Text Normalization with Limited Data for Fast Deployment in Text to Speech Systems ...
|
|
|
|
BASE
|
|
Show details
|
|
19 |
CaSiNo: A Corpus of Campsite Negotiation Dialogues for Automatic Negotiation Systems ...
|
|
|
|
BASE
|
|
Show details
|
|
20 |
multiPRover: Generating Multiple Proofs for Improved Interpretability in Rule Reasoning ...
|
|
|
|
BASE
|
|
Show details
|
|
|
|