The increased ubiquitousness of small smart devices, such as cell- phones, tablets, smart watches and laptops, has led to unique user data, which can be locally processed. The sensors (e.g., microphones and webcam) and improved hardware of the new devices have al- lowed running deep learning models that 20 years ago would have been exclusive to high-end expensive machines. In spite of this progress, state-of-the-art algorithms for facial expression recognition (FER) rely on architectures that cannot be implemented on these devices due to computational and memory constraints. Alternatives involving cloud-based solutions impose privacy barriers that prevent their adoption or user acceptance in wide range of applications. This paper proposes a lightweight model that can run in real-time for image facial expression recognition (IFER) and video facial expression recognition (VFER). The approach relies on a personalization mechanism locally implemented for each subject by fine-tuning a central VFER model with unlabeled videos from a target subject. We train the IFER model to generate pseudo labels and we select the videos with the highest confident predictions to be used for adaptation. The adaptation is performed by implementing a federated learning strategy where the weights of the local model are averaged and used by the central VFER model. We demonstrate that this approach can improve not only the performance on the edge device providing personalized models to the users, but also the central VFER model. We implement a federated learning strategy where the weights of the local models are averaged and used by the central VFER. Within corpus and cross-corpus evaluations on two emotional databases demonstrate that edge models adapted with our personalization strategy achieve up to 13.1% gains in F1-scores. Furthermore, the federated learning implementation improves the mean micro F1-score of the central VFER model by up to 3.4%. The proposed lightweight solution is ideal for interactive user interfaces that preserve the data of the users.
more »
« less
This content will become publicly available on June 24, 2025
Enabling On-Device Large Language Model Personalization with Self-Supervised Data Selection and Synthesis
After a large language model (LLM) is deployed on edge devices, it is desirable for these devices to learn from user-generated conversation data to generate user-specific and personalized responses in real-time. However, user-generated data usually contains sensitive and private information, and uploading such data to the cloud for annotation is not preferred if not prohibited. While it is possible to obtain annotation locally by directly asking users to provide preferred responses, such annotations have to be sparse to not affect user experience. In addition, the storage of edge devices is usually too limited to enable large-scale fine-tuning with full user-generated data. It remains an open question how to enable on-device LLM personalization, considering sparse annotation and limited on-device storage. In this paper, we propose a novel framework to select and store the most representative data online in a self-supervised way. Such data has a small memory footprint and allows infrequent requests of user annotations for further fine-tuning. To enhance fine-tuning quality, multiple semantically similar pairs of question texts and expected responses are generated using the LLM. Our experiments show that the proposed framework achieves the best user-specific content-generating capability (accuracy) and fine-tuning speed (performance) compared with vanilla baselines. To the best of our knowledge, this is the very first on-device LLM personalization framework.
more »
« less
- Award ID(s):
- 2122320
- PAR ID:
- 10548545
- Publisher / Repository:
- ACM/IEEE
- Date Published:
- ISBN:
- 979-8-3503-2348-1
- Format(s):
- Medium: X
- Location:
- San Francisco
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
We study interactive learning of LLM-based language agents based on user edits made to the agent's output. In a typical setting such as writing assistants, the user interacts with a language agent to generate a response given a context, and may optionally edit the agent response to personalize it based on their latent preference, in addition to improving the correctness. The edit feedback is naturally generated, making it a suitable candidate for improving the agent's alignment with the user's preference, and for reducing the cost of user edits over time. We propose a learning framework, PRELUDE that infers a description of the user's latent preference based on historic edit data. The inferred user preference descriptions are used to define prompts for generating responses in the future. This avoids fine-tuning the agent, which is costly, challenging to scale with the number of users, and may even degrade its performance on other tasks. Furthermore, learning descriptive preference improves interpretability, allowing the user to view and modify the learned preference. However, user preference can be complex, subtle, and vary based on context, making it challenging to learn. To address this, we propose a simple yet effective algorithm named CIPHER that leverages the LLM to infer the user preference for a given context based on user edits. In the future, CIPHER retrieves inferred preferences from the k-closest contexts in the history, and forms an aggregate preference for response generation. We introduce two interactive environments -- summarization and email writing, and use a GPT-4 simulated user for evaluation. On both tasks, CIPHER outperforms several baselines by achieving the lowest edit distance cost while only having a small overhead in LLM query cost. Our analysis reports that user preferences learned by CIPHER show significant similarity to the ground truth latent preferences.more » « less
-
The success of ChatGPT is reshaping the landscape of the entire IT industry. The large language model (LLM) powering ChatGPT is experiencing rapid development, marked by enhanced features, improved accuracy, and reduced latency. Due to the execution overhead of LLMs, prevailing commercial LLM products typically manage user queries on remote servers. However, the escalating volume of user queries and the growing complexity of LLMs have led to servers becoming bottlenecks, compromising the quality of service (QoS). To address this challenge, a potential solution is to shift LLM inference services to edge devices, a strategy currently being explored by industry leaders such as Apple, Google, Qualcomm, Samsung, and others. Beyond alleviating the computational strain on servers and enhancing system scalability, deploying LLMs at the edge offers additional advantages. These include real-time responses even in the absence of network connectivity and improved privacy protection for customized or personal LLMs. This article delves into the challenges and potential bottlenecks currently hindering the effective deployment of LLMs on edge devices. Through deploying the LLaMa-2 7B model with INT4 quantization on diverse edge devices and systematically analyzing experimental results, we identify insufficient memory and/or computing resources on traditional edge devices as the primary obstacles. Based on our observation and empirical analysis, we further provide insights and design guidance for the next generation of edge devices and systems from both hardware and software directionsmore » « less
-
more » « less
Abstract Insect pests cause significant damage to food production, so early detection and efficient mitigation strategies are crucial. There is a continual shift toward machine learning (ML)‐based approaches for automating agricultural pest detection. Although supervised learning has achieved remarkable progress in this regard, it is impeded by the need for significant expert involvement in labeling the data used for model training. This makes real‐world applications tedious and oftentimes infeasible. Recently, self‐supervised learning (SSL) approaches have provided a viable alternative to training ML models with minimal annotations. Here, we present an SSL approach to classify 22 insect pests. The framework was assessed on raw and segmented field‐captured images using three different SSL methods, Nearest Neighbor Contrastive Learning of Visual Representations (NNCLR), Bootstrap Your Own Latent, and Barlow Twins. SSL pre‐training was done on ResNet‐18 and ResNet‐50 models using all three SSL methods on the original RGB images and foreground segmented images. The performance of SSL pre‐training methods was evaluated using linear probing of SSL representations and end‐to‐end fine‐tuning approaches. The SSL‐pre‐trained convolutional neural network models were able to perform annotation‐efficient classification. NNCLR was the best performing SSL method for both linear and full model fine‐tuning. With just 5% annotated images, transfer learning with ImageNet initialization obtained 74% accuracy, whereas NNCLR achieved an improved classification accuracy of 79% for end‐to‐end fine‐tuning. Models created using SSL pre‐training consistently performed better, especially under very low annotation, and were robust to object class imbalances. These approaches help overcome annotation bottlenecks and are resource efficient.
-
null (Ed.)Annotated IMU sensor data from smart devices and wearables are essential for developing supervised models for fine-grained human activity recognition, albeit generating sufficient annotated data for diverse human activities under different environments is challenging. Existing approaches primarily use human-in-the-loop based techniques, including active learning; however, they are tedious, costly, and time-consuming. Leveraging the availability of acoustic data from embedded microphones over the data collection devices, in this paper, we propose LASO, a multimodal approach for automated data annotation from acoustic and locomotive information. LASO works over the edge device itself, ensuring that only the annotated IMU data is collected, discarding the acoustic data from the device itself, hence preserving the audio-privacy of the user. In the absence of any pre-existing labeling information, such an auto-annotation is challenging as the IMU data needs to be sessionized for different time-scaled activities in a completely unsupervised manner. We use a change-point detection technique while synchronizing the locomotive information from the IMU data with the acoustic data, and then use pre-trained audio-based activity recognition models for labeling the IMU data while handling the acoustic noises. LASO efficiently annotates IMU data, without any explicit human intervention, with a mean accuracy of 0.93 ($\pm 0.04$) and 0.78 ($\pm 0.05$) for two different real-life datasets from workshop and kitchen environments, respectively.more » « less
