Charting the Growth of Social-Physical HRI (spHRI): A Systematic Review Pipeline Augmented by Small Language Models

2026-06-24 · Source: Artificial Intelligence · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Robotics & Autonomous Systems, Data Science & Analytics · Depth: Advanced, quick

Summary

A systematic review pipeline, augmented by Small Language Models (SLMs), was evaluated for title and abstract screening. This focused on the rapidly growing field of social-physical human-robot interaction (spHRI). SLMs, defined as models under 1.5 billion parameters, did not match human reviewer performance. However, they operated locally and screened papers orders of magnitude faster. A combined SLM ensemble identified 39 papers (10.29% of the final relevant dataset) that human reviewers had missed. These results demonstrate SLMs can augment expert reviewers. This makes large-scale literature reviews more accessible and sustainable in fragmented domains like spHRI.

Key takeaway

For Research Scientists conducting large-scale systematic literature reviews, you should integrate Small Language Models into your screening workflow. While SLMs won't replace human expertise, they significantly accelerate initial title and abstract screening. This augmentation helps identify relevant papers you might otherwise miss, like the 39 papers found in the spHRI review. Implementing local SLMs makes your review process more efficient and comprehensive, ensuring sustainability for extensive literature mapping efforts.

Key insights

SLMs can augment human expert reviewers for faster, more comprehensive large-scale systematic literature reviews.

Principles

• SLMs augment, not replace, human expertise.
• Local SLM operation enables rapid screening.
• Ensemble SLMs improve recall in reviews.

Method

The article describes evaluating SLMs (< 1.5B parameters) for title and abstract screening in spHRI systematic reviews, comparing their speed and recall against human reviewers.

In practice

• Deploy SLMs for initial literature screening.
• Combine SLM outputs with human review.
• Use SLMs to identify missed relevant papers.

Topics

• Social-Physical HRI
• Systematic Review
• Small Language Models
• Literature Screening
• AI Augmentation
• Human-Robot Interaction

Best for: AI Scientist, Research Scientist, Robotics Engineer

Related on AIssential

• Sem-Detect: Semantic Level Detection of AI Generated Peer-Reviews — AI-generated peer reviews converge semantically, while human reviews, even LLM-refined, retain unique ideas.
• Small, Private Language Models as Teammates for Educational Assessment Design — SLMs offer competitive, privacy-sensitive assessment design, but require human oversight due to model evaluation biases.
• Most Influential SIGIR Papers (2026-03 Version) — LLMs and RAG are rapidly advancing information retrieval and recommendation systems, with ongoing innovation in robustness and personalization.
• AISysRev -- LLM-based Tool for Title-abstract Screening — LLMs can significantly expedite systematic review screening, but human oversight remains critical for ambiguous cases.
• SSM Adapters via Hankel Reduced-order Modeling: Injection Site Determines Task Suitability in Long-Context Fine-Tuning — HRM adapters, an SSM-based PEFT method, outperform LoRA for long-context sequential tasks by modulating recurrence via Hankel Reduced-order Modeling.
• Meet an AI Using Scholar — AI tools show promise for systematic review screening and descriptive data extraction, but struggle with comprehensive search and nuanced data interpretation.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Artificial Intelligence.