Speech and Language Technology (SaLT) at the University of Stuttgart

IMS-Toucan

Multilingual and Controllable Text-to-Speech Toolkit of the Speech and Language Technologies Group at the University of Stuttgart.

adviser

ADvISER is a flexible framework to encourage task-oriented dialog system research & development

speaker-anonymization

Speaker anonymization pipeline for hiding the identity of the speaker of a recording by changing the voice in it.

VoicePAT

VoicePAT is a modular and efficient toolkit for voice privacy research, with main focus on speaker anonymization.

cyclegan-emotion-transfer

CycleGAN-based Emotion Style Transfer as Data Augmentation for Speech Emotion Recognition

reading-comprehension

Comparing attention-based convolutional and recurrent neural networks under adversarial attacks to investigate their success and limitations in machine reading comprehension

nlg-eval

Code accompanying the INLG 2018 paper Sequence-to-Sequence Models for Data-to-Text Natural Language Generation: Word- vs. Character-based Processing and Output Diversity

conversational-tree-search

Code and Data for Conversational Tree Search: A new task that bridges the gap between FAQ-style information retrieval and task-oriented dialog.

IMS-Speech

IMS-Speech is a tool for German, English and Russian speech transcription aiming to facilitate research in various disciplines. We are willing to provide a speech transcription service with an intuitive web interface accessible with a wide range of computing devices and to people with various backgrounds. Our service is available here: https://75474978-c3fa-43a5-aa6c-ee36f2513064.ma.bw-cloud-instance.org/ims-speech/

Intrinsic-Subgraph-Generation-for-VQA

hard-negative-captions

diagraph

DIAGRAPH: An open-source graphic interface for dialog flow design

multilingual-seq2seq-slu

Materials for the publication "Leveraging Multilingual Self-Supervised Pretrained Models for Sequence-to-Sequence End-to-End Spoken Language Understanding"

BetterFinetuning

Code accompanying our paper on finetuning self-supervised general speech representations with a combination of contrastive and non-contrastive methods.

bloomzmms

Materials for the publication "Teaching a Multilingual Large Language Model to Understand Multilingual Speech via Multi-Instructional Training"

ethics_in_chatbot_design

A project exploring ethical implications of chatbot design, in particular affective language style. The repository contains code, survey responses, and annotated data for the experiment conducted using this implementation.

Our_Fault

A collaborative dialog game playable by a human and an AI system, designed to better understand how users view such an AI partner. The repository contains code for the game as well as dialog logs, survey responses, and annotations from a user study conducted with this scenario.