Pith. Sign in

RECORD

Prompt Learning for Domain Adaptation in Task-Oriented Dialogue

PaperPDF from arXiv

data files

Paper breakdownThe authors' claims, conditions and measurements, as JSON Signed receiptProof this review has not been altered
X Bluesky LinkedIn Reddit Hacker News
Makesh Narsimhan Sreedhar, Christopher Parisien

Computation and Language · Submitted 10 Nov 2022

abstract compiled · 2026-09-23 · no language-model review

The authors' main claim By tuning soft prompts for a frozen large language model, we show that canonical forms generalize very well to new, unseen domains in a zero- or few-shot setting.

how it is stated · asserted 1.00 read the record →

Comment (0) Read on arXiv ↗
Cite ▾
arXiv ID2211.05596 (version 1)
Pith IDSGDPUA6A, Pith's permanent record of this paper and its review
CitationChristopher Parisien, Makesh Narsimhan Sreedhar. "Prompt Learning for Domain Adaptation in Task-Oriented Dialogue." arXiv:2211.05596 (2022). https://pith.science/paper/SGDPUA6A
Claim paper

The compiled record

The paper's own sentences, arranged. Each label is drawn from a closed vocabulary; the number beside it is the probability the classifier gave that label. Numbers in the tables are the authors', checked by code. No sentence in this record was written by a model.

Open the compiled record

Compiled by classifier-1.13.0 via TypeSafe API, question bank abstract-p1-20260923, 2026-09-23: 1 request, 4,797 input tokens.

What the abstract itself states, as probabilities: claims first 0.04 · states limitation 0.10 · human participants 0.09 · controlled comparison 0.15 · randomized assignment 0.01 · states quantitative result 0.05 · compares with prior results 0.32 · mentions code or data availability 0.03.

The abstract, sentence by sentence

arXiv listing abstract v1. Results and the main claim are underlined; .

Conversation designers continue to face significant obstacles when creating production quality task-oriented dialogue systems. The complexity and cost involved in schema development and data collection is often a major barrier for such designers, limiting their ability to create natural, user-friendly experiences. We frame the classification of user intent as the generation of a canonical form, a lightweight semantic representation using natural language. We show that canonical forms offer a promising alternative to traditional methods for intent classification. By tuning soft prompts for a frozen large language model, we show that canonical forms generalize very well to new, unseen domains in a zero- or few-shot setting. The method is also sample-efficient, reducing the complexity and effort of developing new task-oriented dialogue domains.

background contribution method result interpretation main claim

What the classifier reports for each sentence. The number is the probability it gave the label.
#sentencerolethe authors' own workstates the main conclusionstrength
s0 Conversation designers continue to face significant obstacles when creating production quality task-oriented dialogue systems. background 1.00 0.04 0.02 asserted 1.00
s1 The complexity and cost involved in schema development and data collection is often a major barrier for such designers, limiting their ability to create natural, user-friendly experiences. background 0.99 0.05 0.03 asserted 0.76
s2 We frame the classification of user intent as the generation of a canonical form, a lightweight semantic representation using natural language. contribution 0.58 0.59 0.07 asserted 0.99
s3 We show that canonical forms offer a promising alternative to traditional methods for intent classification. result 0.54 0.88 0.33 asserted 0.82
s4 By tuning soft prompts for a frozen large language model, we show that canonical forms generalize very well to new, unseen domains in a zero- or few-shot setting. result 0.95 0.95 0.59 asserted 1.00
s5 The method is also sample-efficient, reducing the complexity and effort of developing new task-oriented dialogue domains. result 0.61 0.82 0.19 asserted 1.00

What kind of paper the classifier reports this is: method or tool 1.00 · evidence from observation or experiment 0.97.

What was examined, and what was not

sources read
  • arXiv listing abstract v1
not examined
  • full text
  • figures and tables
  • references and citing sentences
  • relations to other papers
levels run
  • P abstract
not run
  • S sections
  • T body sentences
  • C citations
  • R cross-paper relations

Every quoted sentence on this record is the authors'. Every label is from Pith's closed vocabulary, with the probability shown beside it. Every number in the measurements table is the authors', checked by code. Record as JSON

The paper
Original abstract

Conversation designers continue to face significant obstacles when creating production quality task-oriented dialogue systems. The complexity and cost involved in schema development and data collection is often a major barrier for such designers, limiting their ability to create natural, user-friendly experiences. We frame the classification of user intent as the generation of a canonical form, a lightweight semantic representation using natural language. We show that canonical forms offer a promising alternative to traditional methods for intent classification. By tuning soft prompts for a frozen large language model, we show that canonical forms generalize very well to new, unseen domains in a zero- or few-shot setting. The method is also sample-efficient, reducing the complexity and effort of developing new task-oriented dialogue domains.

Human review

How Pith checks papers → · Full audit of this paper →

How the literature uses it

What this paper does with the arXiv papers it cites

Each citing sentence is taken from the citing paper's LaTeX source and typed by Pith's classifier. A label appears at probability 0.8 or more (0.95 for support), a contest at 0.7 or more; the quoted words are the citing authors'. How each label was checked · JSON

What it cites

Uses the method or tool of 2 papers
  • * Using Sentence Transformers [ref] : We use the miniLM-QA [ref] transformer model that has been pretrained on multiple datasets on the text entailment/semantic search task, i.e., given a query and a set of keys (documents/labels), it ranks the keys in order of relevance.

    Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks · Evaluation Method · 1.00

  • * Using Fasttext Embeddings [ref] : We take the mean of all the embedding vectors of the generated canonical form and consider the vector obtained to be the representation of the whole sequence.

    Enriching Word Vectors with Subword Information · Evaluation Method · 1.00

Discussion

Know something about this paper? Leave a comment, or highlight any passage above to comment on it.