🤓 Latest Submissions
Kartik - Adaptive Inference Under Constraints
This project translates recent advances in adaptive model routing, confidence-aware prediction and cost-sensitive inference into a practical system for reliable AI under strict resource constraints. Adaptive Inference Under Constraints investigates how language-model workloads can be executed when compute, memory, time, and external inference budgets are simultaneously limited. Instead of applying one computational strategy to every request, it treats inference as a sequence of decisions made under uncertainty. Development produced practical findings on cascade routing, task-sensitive resource allocation and deadline-aware execution. Experiments showed that inexpensive early routing decisions can often approximate more computationally expensive procedures, while selective verification preserves reliability for uncertain tasks. They also demonstrated that aggressive input compression can reduce answer quality, universal generation limits are poorly suited to diverse workloads, and routing must account for execution time as well as predicted accuracy. The system supports reasoning, factual knowledge, code, information extraction, classification, and summarization. Its execution policy adapts to each request while considering expected quality, computational cost, uncertainty, and remaining runtime. Reliability is treated as part of the inference problem, not as an operational afterthought. Bounded execution, controlled escalation, recovery mechanisms and deadline-aware completion prevent a single expensive request from compromising an entire workload. The broader contribution is a practical demonstration that efficient AI is not achieved simply by choosing a smaller model. It emerges from allocating limited computation intelligently: deciding when local processing is sufficient, when additional capability is justified, and when further computation is no longer worth its cost.
13 Jul 2026
Simplify
The AI-powered Document Review System is designed to transform complex, bureaucratic government content into clear, accessible language that adheres to the Australian Government Style Manual. Government documents, often filled with technical language and jargon, can be overwhelming for many citizens, particularly those from CALD (Culturally and Linguistically Diverse) communities, older Australians, people in rural areas and those with low literacy levels. The system simplifies these documents by automatically applying the rules of the Australian Government Style Manual, improving the clarity and accessibility of government content for all. Key features of the system include AI-driven language simplification, style guideline compliance and accessibility improvements. It rewrites complex bureaucratic language into plain English, ensuring that government documents are both easy to understand and follow. Additionally, it ensures that documents meet Australian English spelling, grammar and tone conventions, which are crucial for consistency in government communication. The system uses multiple specialized AI agents to focus on different aspects of the document review process, including grammar, structure, accessibility, formatting and citations. These agents work together to improve document quality, streamline the revision process and reduce the time required for manual review. This solution is ideal for government agencies looking to enhance communication with citizens and improve public engagement. By leveraging AI, the system automates and accelerates the review process, making it faster, more accurate and efficient. It also helps agencies meet their legal obligations to provide accessible information to all Australians. Demo login: [email protected] Password: trXNiD3WepqPCc3
11 Nov 2024
