Rochester Institute of Technology

Robustness tests for tool-using agents can now check, before the agent runs, whether an edited test case still tests the same task; rescoring the same 14,336 stored trajectories under executable task contracts drops Mistral's attacked-task count from 64 to 26 under a MiniLM cosine threshold.

Robustness tests for tool-using agents can now check, before the agent runs, whether an edited test case still tests the same task; rescoring the same 14,336 stored trajectories under executable task contracts drops Mistral's attacked-task count from 64 to 26 under a MiniLM cosine threshold.

This work extends EXAQC, an evolutionary framework for automated quantum circuit discovery, to image classification by evolving parameterized quantum circuits as intermediate processing modules while retaining classical feature-extraction and prediction layers, achieving 98.42%, 90.62%, and 85.47% accuracy on MNIST, Fashion-MNIST, and CIFAR-10 with gate counts comparable to other quantum architecture-search methods, reaching 85.68% on CIFAR-10 with over 25x fewer trainable parameters, and finding that rotation-based encodings (RX, RY, U3) outperform amplitude encoding by 22-25 points on CIFAR-10.

This work extends EXAQC, an evolutionary framework for automated quantum circuit discovery, to image classification by evolving parameterized quantum circuits as intermediate processing modules while retaining classical feature-extraction and prediction layers, achieving 98.42%, 90.62%, and 85.47% accuracy on MNIST, Fashion-MNIST, and CIFAR-10 with gate counts comparable to other quantum architecture-search methods, reaching 85.68% on CIFAR-10 with over 25x fewer trainable parameters, and finding that rotation-based encodings (RX, RY, U3) outperform amplitude encoding by 22-25 points on CIFAR-10.

The work proposes RapidUn, an influence-guided framework that converts cross-sample influence estimates into fixed sample-specific weights for weighted LoRA unlearning; across Llama-3-8B on Dolly-15k and Alpaca-57k, with cross-model validation on Mistral-7B + Dolly-15k, it achieves lower seen-trigger and OOD-trigger-family ASR than Fisher, GA, and LoReUn while maintaining competitive clean utility, attains a 77x wall-clock speedup over the clean-corpus LoRA retraining reference on Llama-3-8B + Alpaca-57k, and is further supported by TOFU, semantic LLM-judge, and IFEval evaluations.

The work proposes RapidUn, an influence-guided framework that converts cross-sample influence estimates into fixed sample-specific weights for weighted LoRA unlearning; across Llama-3-8B on Dolly-15k and Alpaca-57k, with cross-model validation on Mistral-7B + Dolly-15k, it achieves lower seen-trigger and OOD-trigger-family ASR than Fisher, GA, and LoReUn while maintaining competitive clean utility, attains a 77x wall-clock speedup over the clean-corpus LoRA retraining reference on Llama-3-8B + Alpaca-57k, and is further supported by TOFU, semantic LLM-judge, and IFEval evaluations.