OCCAM: Towards Cost-Efficient and Accuracy-Aware Image Classification Inference

Image classification is a fundamental building block for a majority of computer vision applications. With the growing popularity and capacity of machine learning models, people can easily access trained image classifiers as a service online or offline. However, model use comes with a cost and classifiers of higher capacity usually incur higher inference costs. To harness the respective strengths of different classifiers, we propose a principled approach, OCCAM, to compute the best classifier assignment strategy over image classification queries (termed as the optimal model portfolio) so that the aggregated accuracy is maximized, under user-specified cost budgets. Our approach uses an unbiased and low-variance accuracy estimator and effectively computes the optimal solution by solving an integer linear programming problem. On a variety of real-world datasets, OCCAM achieves 40% cost reduction with little to no accuracy drop.

Hybrid LLM: Cost-Efficient and Quality-Aware Query Routing

Large language models (LLMs) excel in most NLP tasks but also require expensive cloud servers for deployment due to their size, while smaller models that can be deployed on lower cost (e.g., edge) devices, tend to lag behind in terms of response quality. Therefore in this work we propose a hybrid inference approach which combines their respective strengths to save cost and maintain quality. Our approach uses a router that assigns queries to the small or large model based on the predicted query difficulty and the desired quality level. The desired quality level can be tuned dynamically at test time to seamlessly trade quality for cost as per the scenario requirements. In experiments our approach allows us to make up to 40% fewer calls to the large model, with no drop in response quality.

LLM Performance Predictors are good initializers for Architecture Search

Large language models (LLMs) have become an integral component in solving a wide range of NLP tasks. In this work, we explore a novel use case of using LLMs to build performance predictors (PP): models that, given a specific deep neural network architecture, predict its performance on a downstream task. We design PP prompts for LLMs consisting of: (i) role: description of the role assigned to the LLM, (ii) instructions: set of instructions to be followed by the LLM to carry out performance prediction, (iii) hyperparameters: a definition of each architecture-specific hyperparameter and (iv) demonstrations: sample architectures along with their efficiency metrics and 'training from scratch' performance. For machine translation (MT) tasks, we discover that GPT-4 with our PP prompts (LLM-PP) can predict the performance of architecture with a mean absolute error matching the SOTA and a marginal degradation in rank correlation coefficient compared to SOTA performance predictors. Further, we show that the predictions from LLM-PP can be distilled to a small regression model (LLM-Distill-PP). LLM-Distill-PP models surprisingly retain the performance of LLM-PP largely and can be a cost-effective alternative for heavy use cases of performance estimation. Specifically, for neural architecture search (NAS), we propose a Hybrid-Search algorithm for NAS (HS-NAS), which uses LLM-Distill-PP for the initial part of search, resorting to the baseline predictor for rest of the search. We show that HS-NAS performs very similar to SOTA NAS across benchmarks, reduces search hours by 50% roughly, and in some cases, improves latency, GFLOPs, and model size.

On Efficient Approximate Queries over Machine Learning Models

The question of answering queries over ML predictions has been gaining attention in the database community. This question is challenging because the cost of finding high quality answers corresponds to invoking an oracle such as a human expert or an expensive deep neural network model on every single item in the DB and then applying the query. We develop a novel unified framework for approximate query answering by leveraging a proxy to minimize the oracle usage of finding high quality answers for both Precision-Target (PT) and Recall-Target (RT) queries. Our framework uses a judicious combination of invoking the expensive oracle on data samples and applying the cheap proxy on the objects in the DB. It relies on two assumptions. Under the Proxy Quality assumption, proxy quality can be quantified in a probabilistic manner w.r.t. the oracle. This allows us to develop two algorithms: PQA that efficiently finds high quality answers with high probability and no oracle calls, and PQE, a heuristic extension that achieves empirically good performance with a small number of oracle calls. Alternatively, under the Core Set Closure assumption, we develop two algorithms: CSC that efficiently returns high quality answers with high probability and minimal oracle usage, and CSE, which extends it to more general settings. Our extensive experiments on five real-world datasets on both query types, PT and RT, demonstrate that our algorithms outperform the state-of-the-art and achieve high result quality with provable statistical guarantees.