DKT-Sem: Deep Knowledge Tracing with Semantic Embeddings

Element · All elements
Evidence · 2 claims (2 for) · 2 studies (2 design), q2 · 0 of 2 report an effect size · 2 claims rest on one study

Description

DKT-Sem is the article's simpler alternative KT method for dialogues, described as a strong baseline. The authors "slightly modify the deep KT (DKT) model [46] to use semantic embeddings of the textual content in dialogues": S-BERT embeddings of the tutor turn, student turn and averaged KC embeddings feed an LSTM, and KC masteries are predicted via a bilinear projection onto KC embeddings. It is trained with the same objective as LLMKT.

Design Implications

Context

Requirements

• Sentence-BERT embeddings (the article uses the all-mpnet-base-v2 model) of tutor turns, student turns and KC descriptions.
• Turn-level correctness labels mapped to a learnable embedding, as in DKT.

Constraints

• Its advantage over existing KT methods is smaller on the larger MathDial dataset than on CoMTA.

Target Learners

• Students in one-on-one math tutoring dialogues with human or LLM-powered tutors (the article uses the CoMTA and MathDial datasets)

Target Learning Goals

• Estimating student knowledge of math knowledge components (Common Core standards) and predicting student response correctness across dialogue turns

Affordances

• Dialogue Knowledge Tracing Framework

Claims

• DKT-Sem, a DKT variant using semantic text embeddings, performs better than existing KT methods on tutoring dialogues, with a smaller margin on the larger MathDial dataset. [+M]
• Among deep learning KT models, simpler DKT performs comparably to complex architectures under limited dialogue data [+W]

Related Elements

Examples

-

Key Sources

• Alexander Scarlatos, Ryan S. Baker, and Andrew Lan. (2024). Exploring Knowledge Tracing in Tutor-Student Dialogues using LLMs. Published in LAK25: The 15th International Learning Analytics and Knowledge Conference. https://arxiv.org/abs/2409.16490

Page metadata

Field	Value
type	element
id	dkt-sem-semantic-embedding-knowledge-tracing
title	DKT-Sem: Deep Knowledge Tracing with Semantic Embeddings
description	DKT-Sem is the article's simpler alternative KT method for dialogues, described as a strong baseline.
status	review
canonical	True
generated	by claude/unspecified on 2026-09-24
trust	unverified — no one has confirmed this page's content
sources	1 — alexander-scarlatos-2024

Frontmatter is the page's machine-readable record — see the Schema & Guide.