EvoRiri: Interpretable Evolutionary Search for Diagnosing Reward Design Failures in a Bangla Mental Health Assistant
Mathematics · 2026-11-01 · 14 (21) · 3982
DOI: 10.3390/math14213982 ↗At a glance
Journal article in Mathematics, published online on 2026-11-01. Written by Jahnnobi Rahman and 1 co-authors. 26 references. The full text is freely available (open access, CC BY 4.0).
Abstract (original language)
Mental health conversational systems must balance empathy, risk-aware safety, and actionable structure, yet optimising hand-designed proxies for these qualities can expose reward design failures. We present EvoRiri, a template-based Bangla–Banglish–English mental health assistant that uses no language model. Its response policy is an eight-gene mixed continuous–discrete genome (empathy, safety and structure weights, a distress-gating gene, two reserved escalation thresholds, a template selector, and a response-part budget). A genetic algorithm (population 100, 20 generations, tournament selection, and uniform crossover, with 2000 evaluations per run) maximises a scalar fitness that weights rule-based empathy, safety and structure scores (0.40/0.40/0.15) with per-example safety penalties. On 2872 de-identified records rendered as synthetic prompts, the evolved genome raises structural completeness from 0.65 to 1.00 and fitness from 0.556 to 0.703 over a fixed reference genome, with comparable safety. A one-factor ablation attributes this gain to a single structure-weight threshold, and over 30 paired seeds, genetic and equal-budget random search are equivalent on validation prompts. Trace analysis exposed failures that aggregate fitness hid, including question-only collapse, empathy-marker stuffing, and response truncation. The contribution is a diagnostic case study of reward design on a synthetic single-turn corpus, not an optimisation advantage or clinical validation.
Cite
APA
Rahman, J., & Kabir, M. M. J. (2026). EvoRiri: Interpretable Evolutionary Search for Diagnosing Reward Design Failures in a Bangla Mental Health Assistant. Mathematics, 14(21), 3982. https://doi.org/10.3390/math14213982
Vancouver
Rahman J, Kabir MMJ. EvoRiri: Interpretable Evolutionary Search for Diagnosing Reward Design Failures in a Bangla Mental Health Assistant. Mathematics. 2026;14(21):3982. doi:10.3390/math14213982
BibTeX
@article{Rahman2026,
author = {Rahman, Jahnnobi and Kabir, Mir Md. Jahangir},
title = {EvoRiri: Interpretable Evolutionary Search for Diagnosing Reward Design Failures in a Bangla Mental Health Assistant},
journal = {Mathematics},
year = {2026},
volume = {14},
number = {21},
pages = {3982},
doi = {10.3390/math14213982}
}
EUVIMED