Grace Wu | Searcharxiv

arXiv2026

The Sustainability Paradox of Biodegradable Packaging: A Life Cycle Perspective on Chitosan-Based Food Packaging

Chitosan-based nanomaterials are being increasingly explored as sustainable alternatives to petroleum-derived food packaging, yet their environmental performance across the full life cycle remains insufficiently understood. This review critically evaluates these systems from a life cycle perspective and examines how material origin, processing pathways, functional performance, and end-of-life behavior collectively influence sustainability outcomes. Beginning with chitin extraction from crustacean waste, key processing steps, including demineralization, deproteinization, and nanoparticle synthesis, are assessed in terms of chemical intensity, energy demand, and associated emissions. Manufacturing routes, including solvent-based and green synthesis approaches, are compared with those of conventional plastics to identify relative environmental burdens. The use phase is analyzed with respect to antimicrobial functionality, shelf life extension, and potential reductions in food waste. End-of-life pathways, including biodegradation and composting, are evaluated alongside uncertainties related to degradation behavior and nanoparticle fate. By synthesizing these stagewise interactions, this review highlights critical trade-offs that are often overlooked in sustainability narratives and examines whether chitosan-based nanomaterials provide net environmental benefits in food packaging applications.

physics.soc-ph↗

arXiv2025

MAR:Multi-Agent Reflexion Improves Reasoning Abilities in LLMs

LLMs have shown the capacity to improve their performance on reasoning tasks through reflecting on their mistakes, and acting with these reflections in mind. However, continual reflections of the same LLM onto itself exhibit degeneration of thought, where the LLM continues to repeat the same errors again and again even with the knowledge that its wrong. To address this problem, we instead introduce multi-agent with multi-persona debators as the method to generate reflections. Through out extensive experimentation, we've found that the leads to better diversity of in the reflections generated by the llm agent. We demonstrate an accuracy of 47% EM HotPot QA (question answering) and 82.7% on HumanEval (programming), both performances surpassing reflection with a single llm.

cs.AI↗