Intent Discovery
16 papers with code • 3 benchmarks • 3 datasets
Given a set of labelled and unlabelled utterances, the idea is to identify existing (known) intents and potential (new intents) intents. This method can be utilised in conversational system setting.
Benchmarks
These leaderboards are used to track progress in Intent Discovery
| Trend | Dataset | Best Model | Paper | Code | Compare |
|---|---|---|---|---|---|
|
|||||
|
|||||
|
ATIS
SNIPS
Persian-ATIS
Most implemented papers
TEXTOIR: An Integrated and Visualized Platform for Text Open Intent Recognition
It is composed of two main modules: open intent detection and open intent discovery.
Z-BERT-A: a zero-shot Pipeline for Unknown Intent detection
gt4sd/zero-shot-bert-adapters
•
•
In our evaluation, we first analyze the quality of the model after adaptive fine-tuning on known classes.
Intent Mining from past conversations for conversational agent
ajaychatterjee/IntentMining
•
•
COLING 2020
In this paper, we present an intent discovery framework that involves 4 primary steps: Extraction of textual utterances from a conversation using a pre-trained domain agnostic Dialog Act Classifier (Data Extraction), automatic clustering of similar user utterances (Clustering), manual annotation of clusters with an intent label (Labeling) and propagation of intent labels to the utterances from the previous step, which are not mapped to any cluster (Label Propagation); to generate intent training data from raw conversations.
Open Intent Discovery through Unsupervised Semantic Clustering and Dependency Parsing
This paper presents an unsupervised two-stage approach to discover intents and generate meaningful intent labels automatically from a collection of unlabeled utterances in a domain.
Benchmark Data and Evaluation Framework for Intent Discovery Around COVID-19 Vaccine Hesitancy
We use this framework to report baseline intent discovery results over VIRADialogs, that highlight the difficulty of this task.
New Intent Discovery with Pre-training and Contrastive Learning
Existing approaches typically rely on a large amount of labeled utterances and employ pseudo-labeling methods for representation learning and clustering, which are label-intensive, inefficient, and inaccurate.
Generalized Intent Discovery: Learning from Open World Dialogue System
myt517/gid_benchmark
•
•
COLING 2022
Traditional intent classification models are based on a pre-defined intent set and only recognize limited in-domain (IND) intent classes.
Watch the Neighbors: A Unified K-Nearest Neighbor Contrastive Learning Framework for OOD Intent Discovery
For OOD clustering stage, we propose a KCC method to form compact clusters by mining true hard negative samples, which bridges the gap between clustering and representation learning.
A Hybrid Architecture for Out of Domain Intent Detection and Intent Discovery
Makbari1997/VAE-KPCA-HDBSCAN
•
•
On the other side, a labeled dataset is needed to train a model for Intent Detection in task-oriented dialogue systems.
A Clustering Framework for Unsupervised and Semi-supervised New Intent Discovery
thuiar/textoir
•
•
New intent discovery is of great value to natural language processing, allowing for a better understanding of user needs and providing friendly services.
