natural-language-processing-4

Response Generation

3260 papers • 126 benchmarks • 313 datasets

A task where an agent should play the $DE$ role and generate a text to respond to a $P$ message.

(Image credit: Papersgraph)

Benchmarks

These leaderboards are used to track progress in response-generation-4

Trend

Dataset

Best Model

Actions

SIMMC2.0

ArgSciChat

MMConv

Libraries

i

Use these libraries to find response-generation-4 models and implementations

budzianowski/multiwoz

6 papers 834

Datasets

Subtasks

No subtasks available.

Most implemented papers

A Hierarchical Latent Variable Encoder-Decoder Model for Generating Dialogues

Joelle Pineau, Yoshua Bengio, Aaron C. Courville, Alessandro Sordoni, Ryan Lowe, Iulian Serban, Laurent Charlin•Sat Apr 30 2016

A neural network-based generative architecture, with stochastic latent variables that span a variable number of time steps, that improves upon recently proposed models and that the latent variables facilitate both the generation of meaningful, long and diverse responses and maintaining dialogue state is proposed.

1134

Content

PaddlePaddle/Knover

3 papers 673

nouhadziri/THRED

3 papers 111

huggingface/transformers

2 papers 128,631

ruc-nlpir/flashrag

2 papers 893

fangleai/Implicit-LVM

2 papers 56

Comet

ArgSciChat

0

Paper Graph

A Diversity-Promoting Objective Function for Neural Conversation Models

Michel Galley, Jianfeng Gao, W. Dolan, Chris Brockett, Jiwei Li•Sat Oct 10 2015

This work proposes using Maximum Mutual Information (MMI) as the objective function in neural models, and demonstrates that the proposed MMI models produce more diverse, interesting, and appropriate responses, yielding substantive gains in BLEU scores on two conversational datasets and in human evaluations.

2600 0

Paper Graph

Unified Language Model Pre-training for Natural Language Understanding and Generation

Jianfeng Gao, M. Zhou, Li Dong, Furu Wei, Nan Yang, Wenhui Wang, Xiaodong Liu, Yu Wang, H. Hon•Tue May 07 2019

A new Unified pre-trained Language Model (UniLM) that can be fine-tuned for both natural language understanding and generation tasks that compares favorably with BERT on the GLUE benchmark, and the SQuAD 2.0 and CoQA question answering tasks.

1640 0

Paper Graph

MASS: Masked Sequence to Sequence Pre-training for Language Generation

Tao Qin, Tie-Yan Liu, Kaitao Song, Xu Tan, Jianfeng Lu•Mon May 06 2019

This work proposes MAsked Sequence to Sequence pre-training (MASS) for the encoder-decoder based language generation tasks, which achieves the state-of-the-art accuracy on the unsupervised English-French translation, even beating the early attention-based supervised model.

1008 0

Paper Graph

NLG Evaluation Metrics Beyond Correlation Analysis: An Empirical Metric Preference Checklist

Mykola Pechenizkiy, V. Menkovski, Iftitahu Ni'mah, Meng Fang•Sun May 14 2023

It is shown that automatic metrics provide a better guidance than human on discriminating system-level performance in Text Summarization and Controlled Generation tasks, and that multi-aspect human-aligned metric (UniEval) is not necessarily dominant over single-aspects human- aligned metrics (CTC, CtrlEval), and task-agnostic metrics (BLEU, BERTScore), particularly in Controlled generation tasks.

21 0

Paper Graph

Multiresolution Recurrent Neural Networks: An Application to Dialogue Response Generation

Yoshua Bengio, Aaron C. Courville, Kartik Talamadupula, Bowen Zhou, Iulian Serban, Tim Klinger, G. Tesauro•Tue May 31 2016

A new class of models called multiresolution recurrent neural networks, which explicitly model natural language generation at multiple levels of abstraction, are introduced, which outperform competing models by a substantial margin and generate more fluent, relevant and goal-oriented responses.

193 0

Paper Graph

Towards Scalable Multi-domain Conversational Agents: The Schema-Guided Dialogue Dataset

Xiaoxue Zang, Abhinav Rastogi, Srinivas Sunkara, Raghav Gupta, Pranav Khaitan•Wed Sep 11 2019

This work introduces the the Schema-Guided Dialogue (SGD) dataset, containing over 16k multi-domain conversations spanning 16 domains, and presents a schema-guided paradigm for task-oriented dialogue, in which predictions are made over a dynamic set of intents and slots provided as input.

679 0

Paper Graph

Adding a benchmark result helps the community track progress.