Transfer Learning for Finetuning Large Language Models

As the landscape of large language models expands, efficiently finetuning for specific tasks becomes increasingly crucial. At the same time, the landscape of parameter-efficient finetuning methods rapidly expands. Consequently, practitioners face a multitude of complex choices when searching for an optimal finetuning pipeline for large language models. To reduce the complexity for practitioners, we investigate transfer learning for finetuning large language models and aim to transfer knowledge about configurations from related finetuning tasks to a new task. In this work, we transfer learn finetuning by meta-learning performance and cost surrogate models for grey-box meta-optimization from a new meta-dataset. Counter-intuitively, we propose to rely only on transfer learning for new datasets. Thus, we do not use task-specific Bayesian optimization but prioritize knowledge transferred from related tasks over task-specific feedback. We evaluate our method on eight synthetic question-answer datasets and a meta-dataset consisting of 1,800 runs of finetuning Microsoft's Phi-3. Our transfer learning is superior to zero-shot, default finetuning, and meta-optimization baselines. Our results demonstrate the transferability of finetuning to adapt large language models more effectively.

Ensembling Finetuned Language Models for Text Classification

Finetuning is a common practice widespread across different communities to adapt pretrained models to particular tasks. Text classification is one of these tasks for which many pretrained models are available. On the other hand, ensembles of neural networks are typically used to boost performance and provide reliable uncertainty estimates. However, ensembling pretrained models for text classification is not a well-studied avenue. In this paper, we present a metadataset with predictions from five large finetuned models on six datasets, and report results of different ensembling strategies from these predictions. Our results shed light on how ensembling can improve the performance of finetuned text classifiers and incentivize future adoption of ensembles in such tasks.

Large Language Models Engineer Too Many Simple Features For Tabular Data

Tabular machine learning problems often require time-consuming and labor-intensive feature engineering. Recent efforts have focused on using large language models (LLMs) to capitalize on their potential domain knowledge. At the same time, researchers have observed ethically concerning negative biases in other LLM-related use cases, such as text generation. These developments motivated us to investigate whether LLMs exhibit a bias that negatively impacts the performance of feature engineering. While not ethically concerning, such a bias could hinder practitioners from fully utilizing LLMs for automated data science. Therefore, we propose a method to detect potential biases by detecting anomalies in the frequency of operators (e.g., adding two features) suggested by LLMs when engineering new features. Our experiments evaluate the bias of four LLMs, two big frontier and two small open-source models, across 27 tabular datasets. Our results indicate that LLMs are biased toward simple operators, such as addition, and can fail to utilize more complex operators, such as grouping followed by aggregations. Furthermore, the bias can negatively impact the predictive performance when using LLM-generated features. Our results call for mitigating bias when using LLMs for feature engineering.

Dynamic Post-Hoc Neural Ensemblers

Ensemble methods are known for enhancing the accuracy and robustness of machine learning models by combining multiple base learners. However, standard approaches like greedy or random ensembles often fall short, as they assume a constant weight across samples for the ensemble members. This can limit expressiveness and hinder performance when aggregating the ensemble predictions. In this study, we explore employing neural networks as ensemble methods, emphasizing the significance of dynamic ensembling to leverage diverse model predictions adaptively. Motivated by the risk of learning low-diversity ensembles, we propose regularizing the model by randomly dropping base model predictions during the training. We demonstrate this approach lower bounds the diversity within the ensemble, reducing overfitting and improving generalization capabilities. Our experiments showcase that the dynamic neural ensemblers yield competitive results compared to strong baselines in computer vision, natural language processing, and tabular data.

Hardware Aware Ensemble Selection for Balancing Predictive Accuracy and Cost

Automated Machine Learning (AutoML) significantly simplifies the deployment of machine learning models by automating tasks from data preprocessing to model selection to ensembling. AutoML systems for tabular data often employ post hoc ensembling, where multiple models are combined to improve predictive accuracy. This typically results in longer inference times, a major limitation in practical deployments. Addressing this, we introduce a hardware-aware ensemble selection approach that integrates inference time into post hoc ensembling. By leveraging an existing framework for ensemble selection with quality diversity optimization, our method evaluates ensemble candidates for their predictive accuracy and hardware efficiency. This dual focus allows for a balanced consideration of accuracy and operational efficiency. Thus, our approach enables practitioners to choose from a Pareto front of accurate and efficient ensembles. Our evaluation using 83 classification datasets shows that our approach sustains competitive accuracy and can significantly improve ensembles' operational efficiency. The results of this study provide a foundation for extending these principles to additional hardware constraints, setting the stage for the development of more resource-efficient AutoML systems.

Don't Waste Your Time: Early Stopping Cross-Validation

State-of-the-art automated machine learning systems for tabular data often employ cross-validation; ensuring that measured performances generalize to unseen data, or that subsequent ensembling does not overfit. However, using k-fold cross-validation instead of holdout validation drastically increases the computational cost of validating a single configuration. While ensuring better generalization and, by extension, better performance, the additional cost is often prohibitive for effective model selection within a time budget. We aim to make model selection with cross-validation more effective. Therefore, we study early stopping the process of cross-validation during model selection. We investigate the impact of early stopping on random search for two algorithms, MLP and random forest, across 36 classification datasets. We further analyze the impact of the number of folds by considering 3-, 5-, and 10-folds. In addition, we investigate the impact of early stopping with Bayesian optimization instead of random search and also repeated cross-validation. Our exploratory study shows that even a simple-to-understand and easy-to-implement method consistently allows model selection to converge faster; in ~94% of all datasets, on average by ~214%. Moreover, stopping cross-validation enables model selection to explore the search space more exhaustively by considering +167% configurations on average within one hour, while also obtaining better overall performance.

Revealing the Hidden Impact of Top-N Metrics on Optimization in Recommender Systems

The hyperparameters of recommender systems for top-n predictions are typically optimized to enhance the predictive performance of algorithms. Thereby, the optimization algorithm, e.g., grid search or random search, searches for the best hyperparameter configuration according to an optimization-target metric, like nDCG or Precision. In contrast, the optimized algorithm, internally optimizes a different loss function during training, like squared error or cross-entropy. To tackle this discrepancy, recent work focused on generating loss functions better suited for recommender systems. Yet, when evaluating an algorithm using a top-n metric during optimization, another discrepancy between the optimization-target metric and the training loss has so far been ignored. During optimization, the top-n items are selected for computing a top-n metric; ignoring that the top-n items are selected from the recommendations of a model trained with an entirely different loss function. Item recommendations suitable for optimization-target metrics could be outside the top-n recommended items; hiddenly impacting the optimization performance. Therefore, we were motivated to analyze whether the top-n items are optimal for optimization-target top-n metrics. In pursuit of an answer, we exhaustively evaluate the predictive performance of 250 selection strategies besides selecting the top-n. We extensively evaluate each selection strategy over twelve implicit feedback and eight explicit feedback data sets with eleven recommender systems algorithms. Our results show that there exist selection strategies other than top-n that increase predictive performance for various algorithms and recommendation domains. However, the performance of the top ~43% of selection strategies is not significantly different. We discuss the impact of our findings on optimization and re-ranking in recommender systems and feasible solutions.

Q(D)O-ES: Population-based Quality (Diversity) Optimisation for Post Hoc Ensemble Selection in AutoML

Automated machine learning (AutoML) systems commonly ensemble models post hoc to improve predictive performance, typically via greedy ensemble selection (GES). However, we believe that GES may not always be optimal, as it performs a simple deterministic greedy search. In this work, we introduce two novel population-based ensemble selection methods, QO-ES and QDO-ES, and compare them to GES. While QO-ES optimises solely for predictive performance, QDO-ES also considers the diversity of ensembles within the population, maintaining a diverse set of well-performing ensembles during optimisation based on ideas of quality diversity optimisation. The methods are evaluated using 71 classification datasets from the AutoML benchmark, demonstrating that QO-ES and QDO-ES often outrank GES, albeit only statistically significant on validation data. Our results further suggest that diversity can be beneficial for post hoc ensembling but also increases the risk of overfitting.

CMA-ES for Post Hoc Ensembling in AutoML: A Great Success and Salvageable Failure

Many state-of-the-art automated machine learning (AutoML) systems use greedy ensemble selection (GES) by Caruana et al. (2004) to ensemble models found during model selection post hoc. Thereby, boosting predictive performance and likely following Auto-Sklearn 1's insight that alternatives, like stacking or gradient-free numerical optimization, overfit. Overfitting in Auto-Sklearn 1 is much more likely than in other AutoML systems because it uses only low-quality validation data for post hoc ensembling. Therefore, we were motivated to analyze whether Auto-Sklearn 1's insight holds true for systems with higher-quality validation data. Consequently, we compared the performance of covariance matrix adaptation evolution strategy (CMA-ES), state-of-the-art gradient-free numerical optimization, to GES on the 71 classification datasets from the AutoML benchmark for AutoGluon. We found that Auto-Sklearn's insight depends on the chosen metric. For the metric ROC AUC, CMA-ES overfits drastically and is outperformed by GES -- statistically significantly for multi-class classification. For the metric balanced accuracy, CMA-ES does not overfit and outperforms GES significantly. Motivated by the successful application of CMA-ES for balanced accuracy, we explored methods to stop CMA-ES from overfitting for ROC AUC. We propose a method to normalize the weights produced by CMA-ES, inspired by GES, that avoids overfitting for CMA-ES and makes CMA-ES perform better than or similar to GES for ROC AUC.

Assembled-OpenML: Creating Efficient Benchmarks for Ensembles in AutoML with OpenML

Automated Machine Learning (AutoML) frameworks regularly use ensembles. Developers need to compare different ensemble techniques to select appropriate techniques for an AutoML framework from the many potential techniques. So far, the comparison of ensemble techniques is often computationally expensive, because many base models must be trained and evaluated one or multiple times. Therefore, we present Assembled-OpenML. Assembled-OpenML is a Python tool, which builds meta-datasets for ensembles using OpenML. A meta-dataset, called Metatask, consists of the data of an OpenML task, the task's dataset, and prediction data from model evaluations for the task. We can make the comparison of ensemble techniques computationally cheaper by using the predictions stored in a metatask instead of training and evaluating base models. To introduce Assembled-OpenML, we describe the first version of our tool. Moreover, we present an example of using Assembled-OpenML to compare a set of ensemble techniques. For this example comparison, we built a benchmark using Assembled-OpenML and implemented ensemble techniques expecting predictions instead of base models as input. In our example comparison, we gathered the prediction data of $1523$ base models for $31$ datasets. Obtaining the prediction data for all base models using Assembled-OpenML took ${\sim} 1$ hour in total. In comparison, obtaining the prediction data by training and evaluating just one base model on the most computationally expensive dataset took ${\sim} 37$ minutes.