JUHE API Marketplace
DATASET
Open Source Community

spoken-language-understanding-research-datasets

This dataset comprises multiple sub‑datasets for speech language understanding research, including data from the SmartLights and SmartSpeaker assistants. The SmartLights dataset is used for cross‑validation and contains six intents for controlling light switches, brightness, or color changes. The SmartSpeaker dataset is used for training/testing, includes English and French versions, and is intended for controlling playback and music on smart speakers.

Updated 1/16/2024
github

Description

Dataset Overview

This dataset is associated with the publication titled “Spoken Language Understanding on the Edge”, accepted in 2019 at the 5th Workshop on Energy Efficient Machine Learning and Cognitive Computing held together with NeurIPS 2019. The dataset aims to promote reproducibility and practical utility for the SLU community, containing thousands of textual queries with intents and slots, together with corresponding speech recordings.

Dataset Description

  1. SmartLights assistant:

    • Used for cross‑validation, contains six intents related to switching lights on/off, adjusting brightness, and changing color.
    • Vocabulary size ≈ 400 words.
    • Specific intents include:
      • DecreaseBrightness (296 queries, slot: room)
      • IncreaseBrightness (296 queries, slot: room)
      • SetLightBrightness (296 queries, slots: room, brightness)
      • SetLightColor (300 queries, slots: room, color)
      • SwitchLightOff (299 queries, slot: room)
      • SwitchLightOn (278 queries, slot: room)
  2. SmartSpeaker assistant:

    • English and French versions, used for training/testing.
    • Training set contains nine intents (eight in French) for controlling a smart speaker, such as volume adjustment and music playback.
    • English vocabulary > 65 k words; French > 70 k words.
    • English intents include:
      • NextSong (200 queries, no slot)
      • PreviousSong (199 queries, no slot)
      • SpeakerInterrupt (172 queries, no slot)
      • ResumeMusic (200 queries, no slot)
      • VolumeDown (215 queries, slot: volume_level_absolute)
      • VolumeUp (260 queries, slot: volume_level_absolute)
      • VolumeSet (100 queries, slots: volume_level_absolute, volume_level_percent)
      • GetInfos (199 queries, slot: music_item)
      • PlayMusic (1508 queries, slots: song_name, artist_name, album_name, playlist_mode, playlist_name)
    • French intents include:
      • NextSong (126 queries, no slot)
      • PreviousSong (62 queries, no slot)
      • SpeakerInterrupt (421 queries, no slot)
      • ResumeMusic (107 queries, no slot)
      • VolumeShift (437 queries, slot: volume_action)
      • VolumeSet (229 queries, slots: volume_level_absolute, volume_level_percent, volume_level_relative)
      • GetInfos (62 queries, no slot)
      • PlayMusic (548 queries in train, 1500 queries in test, slots: song_name, artist_name, album_name, playlist_mode, playlist_name)

Dataset Access

Access requires a request through this form; permission will be granted promptly.

License Summary

The dataset is limited to academic and/or research use; commercial use is prohibited. Redistributions must retain the unmodified dataset and follow the same license terms. Any publication must include a full citation of the original paper.

AI studio

Generate PPTs instantly with Nano Banana Pro.

Generate PPT Now

Access Dataset

Login to Access

Please login to view download links and access full dataset details.

Topics

Speech Recognition
Smart Home

Source

Organization: github

Created: 10/17/2018

Power Your Data Analysis with Premium AI Models

Supporting GPT-5, Claude-4, DeepSeek v3, Gemini and more.

Enjoy a free trial and save 20%+ compared to official pricing.