1 |
GreaseLM: Graph REASoning Enhanced Language Models for Question Answering ...
|
|
|
|
BASE
|
|
Show details
|
|
2 |
Prefix-Tuning: Optimizing Continuous Prompts for Generation ...
|
|
|
|
BASE
|
|
Show details
|
|
4 |
Conditional probing: measuring usable information beyond a baseline ...
|
|
|
|
BASE
|
|
Show details
|
|
5 |
Swords: A Benchmark for Lexical Substitution with Improved Data Coverage and Quality ...
|
|
|
|
BASE
|
|
Show details
|
|
6 |
Swords: A Benchmark for Lexical Substitution with Improved Data Coverage and Quality ...
|
|
|
|
BASE
|
|
Show details
|
|
7 |
Designing and Interpreting Probes with Control Tasks ...
|
|
|
|
Abstract:
Probes, supervised models trained to predict properties (like parts-of-speech) from representations (like ELMo), have achieved high accuracy on a range of linguistic tasks. But does this mean that the representations encode linguistic structure or just that the probe has learned the linguistic task? In this paper, we propose control tasks, which associate word types with random outputs, to complement linguistic tasks. By construction, these tasks can only be learned by the probe itself. So a good probe, (one that reflects the representation), should be selective, achieving high linguistic task accuracy and low control task accuracy. The selectivity of a probe puts linguistic task accuracy in context with the probe's capacity to memorize from word types. We construct control tasks for English part-of-speech tagging and dependency edge prediction, and show that popular probes on ELMo representations are not selective. We also find that dropout, commonly used to control probe complexity, is ineffective for ... : EMNLP 2019 ...
|
|
Keyword:
Computation and Language cs.CL; FOS Computer and information sciences
|
|
URL: https://dx.doi.org/10.48550/arxiv.1909.03368 https://arxiv.org/abs/1909.03368
|
|
BASE
|
|
Hide details
|
|
8 |
Unanimous Prediction for 100\% Precision with Application to Learning Semantic Mappings
|
|
|
|
In: arXiv (2019)
|
|
BASE
|
|
Show details
|
|
10 |
From Language to Programs: Bridging Reinforcement Learning and Maximum Marginal Likelihood ...
|
|
|
|
BASE
|
|
Show details
|
|
11 |
Learning Symmetric Collaborative Dialogue Agents with Dynamic Knowledge Graph Embeddings ...
|
|
|
|
BASE
|
|
Show details
|
|
13 |
Learning Executable Semantic Parsers for Natural Language Understanding ...
|
|
|
|
BASE
|
|
Show details
|
|
15 |
Learning Dependency-Based Compositional Semantics
|
|
Liang, Percy. - : eScholarship, University of California, 2011
|
|
In: Liang, Percy. (2011). Learning Dependency-Based Compositional Semantics. UC Berkeley: Computer Science. Retrieved from: http://www.escholarship.org/uc/item/1b1189cm (2011)
|
|
BASE
|
|
Show details
|
|
|
|