We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.CL

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Computation and Language

Title: GenNI: Human-AI Collaboration for Data-Backed Text Generation

Abstract: Table2Text systems generate textual output based on structured data utilizing machine learning. These systems are essential for fluent natural language interfaces in tools such as virtual assistants; however, left to generate freely these ML systems often produce misleading or unexpected outputs. GenNI (Generation Negotiation Interface) is an interactive visual system for high-level human-AI collaboration in producing descriptive text. The tool utilizes a deep learning model designed with explicit control states. These controls allow users to globally constrain model generations, without sacrificing the representation power of the deep learning models. The visual interface makes it possible for users to interact with AI systems following a Refine-Forecast paradigm to ensure that the generation system acts in a manner human users find suitable. We report multiple use cases on two experiments that improve over uncontrolled generation approaches, while at the same time providing fine-grained control. A demo and source code are available at this https URL .
Comments: IEEE VIS 2021
Subjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Human-Computer Interaction (cs.HC)
ACM classes: I.2.7; H.5.2
DOI: 10.1109/TVCG.2021.3114845
Cite as: arXiv:2110.10185 [cs.CL]
  (or arXiv:2110.10185v1 [cs.CL] for this version)

Submission history

From: Hendrik Strobelt [view email]
[v1] Tue, 19 Oct 2021 18:07:07 GMT (2558kb,D)

Link back to: arXiv, form interface, contact.