From Instructions to Assistance: a Dataset Aligning Instruction Manuals with Assembly Videos for Evaluating Multimodal LLMs

TL;DR

This paper introduces a dataset linking instruction manuals with assembly videos to evaluate multimodal large language models' ability to assist in technical tasks, focusing on reasoning, step tracking, and manual referencing.

Contribution

It presents the Manual to Action Dataset (M2AD) for assessing MLMs in procedural understanding and introduces evaluation benchmarks for reasoning and referencing capabilities.

Findings

Some MLMs understand procedural sequences but are limited by architecture.

Performance is constrained by hardware and architectural factors.

Multi-image and interleaved text-image reasoning are needed for improvement.

Abstract

The recent advancements introduced by Large Language Models (LLMs) have transformed how Artificial Intelligence (AI) can support complex, real world tasks, pushing research outside the text boundaries towards multi modal contexts and leading to Multimodal Large Language Models (MLMs). Given the current adoption of LLM based assistants in solving technical or domain specific problems, the natural continuation of this trend is to extend the input domains of these assistants exploiting MLMs. Ideally, these MLMs should be used as real time assistants in procedural tasks, hopefully integrating a view of the environment where the user being assisted is, or even better sharing the same point of view via Virtual Reality (VR) or Augmented Reality (AR) supports, to reason over the same scenario the user is experiencing. With this work, we aim at evaluating the quality of currently openly available MLMs to provide this kind of assistance on technical tasks. To this end, we annotated a data set of furniture assembly with step by step labels and manual references: the Manual to Action Dataset (M2AD). We used this dataset to assess (1) to which extent the reasoning abilities of MLMs can be used to reduce the need for detailed labelling, allowing for more efficient, cost effective annotation practices, (2) whether MLMs are able to track the progression of assembly steps (3) and whether MLMs can refer correctly to the instruction manual pages. Our results showed that while some models understand procedural sequences, their performance is limited by architectural and hardware constraints, highlighting the need for multi image and interleaved text image reasoning.