Foundation-Preserving Adaptation via Generalized Rayleigh-Quotient Optimization

While finetuning effectively adapts foundation models to specialized downstream tasks, it can degrade nontarget capabilities acquired during pretraining. Existing forgetting aware methods typically seek safer updates through specialized initialization or fixed constraints, but do not regulate the adaptation preservation trade-off during training. We propose Foundation Preserving LoRA (FoLoRA), a forgetting aware optimization framework. Guided by a first order preservation condition, FoLoRA defines a forgetting penalty over pretraining-proxy activations and a task utility over downstream task activations. It then scores update directions by task utility per unit forgetting penalty via a generalized Rayleigh quotient. The resulting spectral coordinate system enables direction wise gated Adam updates, attenuating low utility to penalty directions during training. To estimate the forgetting penalty, FoLoRA constructs pretraining proxy calibration data by sampling from the pretrained model rather than relying on a single proxy dataset. Experiments on math, code, and instruction following adaptation show that FoLoRA achieves the strongest preservation adaptation balance over baselines, improving target task performance with best aggregate preservation of non target capabilities.

Importance Sampling via Score-based Generative Models

Importance sampling, which involves sampling from a probability density function (PDF) proportional to the product of an importance weight function and a base PDF, is a powerful technique with applications in variance reduction, biased or customized sampling, data augmentation, and beyond. Inspired by the growing availability of score-based generative models (SGMs), we propose an entirely training-free Importance sampling framework that relies solely on an SGM for the base PDF. Our key innovation is realizing the importance sampling process as a backward diffusion process, expressed in terms of the score function of the base PDF and the specified importance weight function--both readily available--eliminating the need for any additional training. We conduct a thorough analysis demonstrating the method's scalability and effectiveness across diverse datasets and tasks, including importance sampling for industrial and natural images with neural importance weight functions. The training-free aspect of our method is particularly compelling in real-world scenarios where a single base distribution underlies multiple biased sampling tasks, each requiring a different importance weight function. To the best of our knowledge our approach is the first importance sampling framework to achieve this.

Learning RL-Policies for Joint Beamforming Without Exploration: A Batch Constrained Off-Policy Approach

In this project, we consider the problem of network parameter optimization for rate maximization. We frame this as a joint optimization problem of power control, beam forming, and interference cancellation. We consider the setting where multiple Base Stations (BSs) are communicating with multiple user equipments (UEs). Because of the exponential computational complexity of brute force search, we instead solve this non-convex optimization problem using deep reinforcement learning (RL) techniques. The modern communication systems are notorious for their difficulty in exactly modeling their behaviour. This limits us in using RL based algorithms as interaction with the environment is needed for the agent to explore and learn efficiently. Further, it is ill advised to deploy the algorithm in real world for exploration and learning because of the high cost of failure. In contrast to the previous RL-based solutions proposed, such as deep-Q network (DQN) based control, we propose taking an offline model based approach. We specifically consider discrete batch constrained deep Q-learning (BCQ) and show that performance similar to DQN can be acheived with only a fraction of the data and without the need for exploration. This results in maximizing sample efficiency and minimizing risk in the deployment of a new algorithm to commercial networks. We provide the entire resource of the project, including code and data, at the following link: https://github.com/Heasung-Kim/ safe-rl-deployment-for-5g.