Optimization Instability in Autonomous Agentic Workflows for Clinical Symptom Detection

TL;DR

This work evaluates optimization instability, a phenomenon in which continued autonomous improvement paradoxically degrades classifier performance, using Pythia, an open-source framework for automated prompt optimization, and demonstrates that retrospective selection outperforms active intervention for stabilization in low-prevalence classification tasks.

Abstract

Autonomous agentic workflows that iteratively refine their own behavior hold considerable promise, yet their failure modes remain poorly characterized. We investigate optimization instability, a phenomenon in which continued autonomous improvement paradoxically degrades classifier performance, using Pythia, an open-source framework for automated prompt optimization. Evaluating three clinical symptoms with varying prevalence (shortness of breath at 23%, chest pain at 12%, and Long COVID brain fog at 3%), we observed that validation sensitivity oscillated between 1.0 and 0.0 across iterations, with severity inversely proportional to class prevalence. At 3% prevalence, the system achieved 95% accuracy while detecting zero positive cases, a failure mode obscured by standard evaluation metrics. We evaluated two interventions: a guiding agent that actively redirected optimization, amplifying overfitting rather than correcting it, and a selector agent that retrospectively identified the best-performing iteration successfully prevented catastrophic failure. With selector agent oversight, the system outperformed expert-curated lexicons on brain fog detection by 331% (F1) and chest pain by 7%, despite requiring only a single natural language term as input. These findings characterize a critical failure mode of autonomous AI systems and demonstrate that retrospective selection outperforms active intervention for stabilization in low-prevalence classification tasks.

Figures (2)

• Figure 1: Pythia multi-agent architecture. The optimization loop begins with a clinical note and prompt $P_t$ passed to the Specialist Agent, which produces binary classification $\hat{y} \in \{0,1\}$. The system computes sensitivity ($\sigma_t$), specificity ($\tau_t$), and F1 score. If the convergence criteria are met, the system outputs the final prompt $P^*$. Otherwise, Error Analysis Agents (orange) generate natural language critiques for false negatives ($\mathcal{FN}$) or false positives ($\mathcal{FP}$), which Synthesis Agents (green) aggregate into a refined prompt $P_{t+1}$. The loop iterates until convergence or reaches the maximum number of iterations.
• Figure 2: Optimization dynamics and comparative performance.(A) Validation sensitivity across seven optimization iterations. Brain fog (3% prevalence) exhibited catastrophic collapse, reaching zero sensitivity at iterations 3 and 6 (marked with X). Higher prevalence conditions showed oscillation but avoided complete failure. (B) Validation F1 scores comparing Pythia with selector agent versus expert curated lexicon baselines. Pythia outperformed lexicons on brain fog (+331%) and chest pain (+7%), while achieving comparable performance on SOB ($-$5%), despite requiring only a single natural-language term as input.