Abstract:Reinforcement learning (RL) algorithms can be divided into two classes: model-free algorithms, which are sample-inefficient, and model-based algorithms, which suffer from model bias. Dyna-style algorithms combine these two approaches by using simulated data from an estimated environmental model to accelerate model-free training. However, their efficiency is compromised when the estimated model is inaccurate. Previous works address this issue by using model ensembles or pretraining the estimated model with data collected from the real environment, increasing computational and sample complexity. To tackle this issue, we introduce an out-of-distribution (OOD) data filter that removes simulated data from the estimated model that significantly diverges from data collected in the real environment. We show theoretically that this technique enhances the quality of simulated data. With the help of the OOD data filter, the data simulated from the estimated model better mimics the data collected by interacting with the real model. This improvement is evident in the critic updates compared to using the simulated data without the OOD data filter. Our experiment integrates the data filter into the model-based policy optimization (MBPO) algorithm. The results demonstrate that our method requires fewer interactions with the real environment to achieve a higher level of optimality than MBPO, even without a model ensemble.
Abstract:Solving for the complex conditions of materials synthesis and processing requires analyzing information gathered from multiple modes of characterization. Currently, quantitative information is extracted serially with manual tools and intuition, constraining the feedback cycle for process optimization. We use machine learning to automate and generalize feature extraction for in-situ reflection high-energy electron diffraction (RHEED) data to establish quantitatively predictive relationships in small sets ($\sim$10) of expert-labeled data, and apply these to save significant time on subsequent epitaxially grown samples. The fidelity of these relationships is tested on a representative material system ($W_{1-x}V_xSe2$ growth on c-plane sapphire substrate (0001)) at two stages of synthesis with two aims: 1) predicting the grain alignment of the deposited film from the pre-growth substrate surface data, and 2) estimating the vanadium (V) dopant concentration using in-situ RHEED as a proxy for ex-situ methods (e.g. x-ray photoelectron spectroscopy). Both tasks are accomplished using the same set of materials agnostic core features, eliminating the need to retrain for specific systems and leading to a potential 80\% time saving over a 100 sample synthesis campaign. These predictions provide guidance for recipe adjustments to avoid doomed trials, reduce follow-on characterization, and improve control resolution for materials synthesis, ultimately accelerating materials discovery and commercial scale-up.
Abstract:Amidst the rise of Large Multimodal Models (LMMs) and their widespread application in generating and interpreting complex content, the risk of propagating biased and harmful memes remains significant. Current safety measures often fail to detect subtly integrated hateful content within ``Confounder Memes''. To address this, we introduce \textsc{HateSieve}, a new framework designed to enhance the detection and segmentation of hateful elements in memes. \textsc{HateSieve} features a novel Contrastive Meme Generator that creates semantically paired memes, a customized triplet dataset for contrastive learning, and an Image-Text Alignment module that produces context-aware embeddings for accurate meme segmentation. Empirical experiments on the Hateful Meme Dataset show that \textsc{HateSieve} not only surpasses existing LMMs in performance with fewer trainable parameters but also offers a robust mechanism for precisely identifying and isolating hateful content. \textcolor{red}{Caution: Contains academic discussions of hate speech; viewer discretion advised.}
Abstract:Prompt tuning provides an efficient way for users to customize Large Language Models (LLMs) with their private data in the emerging LLM service scenario. However, the sensitive nature of private data brings the need for privacy preservation in LLM service customization. Based on prompt tuning, we propose Privacy-Preserving Prompt Tuning (RAPT), a framework that provides privacy guarantees for LLM services. \textsc{rapt} adopts a local privacy setting, allowing users to privatize their data locally with local differential privacy. As prompt tuning performs poorly when directly trained on privatized data, we introduce a novel privatized token reconstruction task that is trained jointly with the downstream task, allowing LLMs to learn better task-dependent representations. Despite the simplicity of our framework, experiments show that RAPT achieves competitive performance across tasks while providing privacy guarantees against adversaries.
Abstract:Echo state network (ESN), a kind of recurrent neural networks, consists of a fixed reservoir in which neurons are connected randomly and recursively and obtains the desired output only by training output connection weights. First-order reduced and controlled error (FORCE) learning is an online supervised training approach that can change the chaotic activity of ESNs into specified activity patterns. This paper proposes a composite FORCE learning method based on recursive least squares to train ESNs whose initial activity is spontaneously chaotic, where a composite learning technique featured by dynamic regressor extension and memory data exploitation is applied to enhance parameter convergence. The proposed method is applied to a benchmark problem about predicting chaotic time series generated by the Mackey-Glass system, and numerical results have shown that it significantly improves learning and prediction performances compared with existing methods.