Bilevel optimization recently has received tremendous attention due to its great success in solving important machine learning problems like meta learning, reinforcement learning, and hyperparameter optimization. Extending single-agent training on bilevel problems to the decentralized setting is a natural generalization, and there has been a flurry of work studying decentralized bilevel optimization algorithms. However, it remains unknown how to design the distributed algorithm with sample complexity and convergence rate comparable to SGD for stochastic optimization, and at the same time without directly computing the exact Hessian or Jacobian matrices. In this paper we propose such an algorithm. More specifically, we propose a novel decentralized stochastic bilevel optimization (DSBO) algorithm that only requires first order stochastic oracle, Hessian-vector product and Jacobian-vector product oracle. The sample complexity of our algorithm matches the currently best known results for DSBO, and the advantage of our algorithm is that it does not require estimating the full Hessian and Jacobian matrices, thereby having improved per-iteration complexity.
The alfalfa crop is globally important as livestock feed, so highly efficient planting and harvesting could benefit many industries, especially as the global climate changes and traditional methods become less accurate. Recent work using machine learning (ML) to predict yields for alfalfa and other crops has shown promise. Previous efforts used remote sensing, weather, planting, and soil data to train machine learning models for yield prediction. However, while remote sensing works well, the models require large amounts of data and cannot make predictions until the harvesting season begins. Using weather and planting data from alfalfa variety trials in Kentucky and Georgia, our previous work compared feature selection techniques to find the best technique and best feature set. In this work, we trained a variety of machine learning models, using cross validation for hyperparameter optimization, to predict biomass yields, and we showed better accuracy than similar work that employed more complex techniques. Our best individual model was a random forest with a mean absolute error of 0.081 tons/acre and R{$^2$} of 0.941. Next, we expanded this dataset to include Wisconsin and Mississippi, and we repeated our experiments, obtaining a higher best R{$^2$} of 0.982 with a regression tree. We then isolated our testing datasets by state to explore this problem's eligibility for domain adaptation (DA), as we trained on multiple source states and tested on one target state. This Trivial DA (TDA) approach leaves plenty of room for improvement through exploring more complex DA techniques in forthcoming work.
AnalogVNN, a simulation framework built on PyTorch which can simulate the effects of optoelectronic noise, limited precision, and signal normalization present in photonic neural network accelerators. We use this framework to train and optimize linear and convolutional neural networks with up to 9 layers and ~1.7 million parameters, while gaining insights into how normalization, activation function, reduced precision, and noise influence accuracy in analog photonic neural networks. By following the same layer structure design present in PyTorch, the AnalogVNN framework allows users to convert most digital neural network models to their analog counterparts with just a few lines of code, taking full advantage of the open-source optimization, deep learning, and GPU acceleration libraries available through PyTorch. Code is available at https://analogvnn.github.io
Temperature is a widely used hyperparameter in various tasks involving neural networks, such as classification or metric learning, whose choice can have a direct impact on the model performance. Most of existing works select its value using hyperparameter optimization methods requiring several runs to find the optimal value. We propose to analyze the impact of temperature on classification tasks by describing a dataset as a set of statistics computed on representations on which we can build a heuristic giving us a default value of temperature. We study the correlation between these extracted statistics and the observed optimal temperatures. This preliminary study on more than a hundred combinations of different datasets and features extractors highlights promising results towards the construction of a general heuristic for temperature.
Gliomas are the most frequent primary brain tumors in adults. Glioma change detection aims at finding the relevant parts of the image that change over time. Although Deep Learning (DL) shows promising performances in similar change detection tasks, the creation of large annotated datasets represents a major bottleneck for supervised DL applications in radiology. To overcome this, we propose a combined use of weak labels (imprecise, but fast-to-create annotations) and Transfer Learning (TL). Specifically, we explore inductive TL, where source and target domains are identical, but tasks are different due to a label shift: our target labels are created manually by three radiologists, whereas our source weak labels are generated automatically from radiology reports via NLP. We frame knowledge transfer as hyperparameter optimization, thus avoiding heuristic choices that are frequent in related works. We investigate the relationship between model size and TL, comparing a low-capacity VGG with a higher-capacity ResNeXt model. We evaluate our models on 1693 T2-weighted magnetic resonance imaging difference maps created from 183 patients, by classifying them into stable or unstable according to tumor evolution. The weak labels extracted from radiology reports allowed us to increase dataset size more than 3-fold, and improve VGG classification results from 75% to 82% AUC. Mixed training from scratch led to higher performance than fine-tuning or feature extraction. To assess generalizability, we ran inference on an open dataset (BraTS-2015: 15 patients, 51 difference maps), reaching up to 76% AUC. Overall, results suggest that medical imaging problems may benefit from smaller models and different TL strategies with respect to computer vision datasets, and that report-generated weak labels are effective in improving model performances. Code, in-house dataset and BraTS labels are released.
In this work we consider a variant of adversarial online learning where in each round one picks $B$ out of $N$ arms and incurs cost equal to the $\textit{minimum}$ of the costs of each arm chosen. We propose an algorithm called Follow the Perturbed Multiple Leaders (FPML) for this problem, which we show (by adapting the techniques of Kalai and Vempala [2005]) achieves expected regret $\mathcal{O}(T^{\frac{1}{B+1}}\ln(N)^{\frac{B}{B+1}})$ over time horizon $T$ relative to the $\textit{single}$ best arm in hindsight. This introduces a trade-off between the budget $B$ and the single-best-arm regret, and we proceed to investigate several applications of this trade-off. First, we observe that algorithms which use standard regret minimizers as subroutines can sometimes be adapted by replacing these subroutines with FPML, and we use this to generalize existing algorithms for Online Submodular Function Maximization [Streeter and Golovin, 2008] in both the full feedback and semi-bandit feedback settings. Next, we empirically evaluate our new algorithms on an online black-box hyperparameter optimization problem. Finally, we show how FPML can lead to new algorithms for Linear Programming which require stronger oracles at the benefit of fewer oracle calls.
In recent years, Artificial Neural Networks (ANNs) have been introduced in Structural Health Monitoring (SHM) systems. A semi-supervised method with a data-driven approach allows the ANN training on data acquired from an undamaged structural condition to detect structural damages. In standard approaches, after the training stage, a decision rule is manually defined to detect anomalous data. However, this process could be made automatic using machine learning methods, whom performances are maximised using hyperparameter optimization techniques. The paper proposes a semi-supervised method with a data-driven approach to detect structural anomalies. The methodology consists of: (i) a Variational Autoencoder (VAE) to approximate undamaged data distribution and (ii) a One-Class Support Vector Machine (OC-SVM) to discriminate different health conditions using damage sensitive features extracted from VAE's signal reconstruction. The method is applied to a scale steel structure that was tested in nine damage's scenarios by IASC-ASCE Structural Health Monitoring Task Group.
As machine learning permeates more industries and models become more expensive and time consuming to train, the need for efficient automated hyperparameter optimization (HPO) has never been more pressing. Multi-step planning based approaches to hyperparameter optimization promise improved efficiency over myopic alternatives by more effectively balancing out exploration and exploitation. However, the potential of these approaches has not been fully realized due to their technical complexity and computational intensity. In this work, we leverage recent advances in Transformer-based, natural-language-interfaced hyperparameter optimization to circumvent these barriers. We build on top of the recently proposed OptFormer which casts both hyperparameter suggestion and target function approximation as autoregressive generation thus making planning via rollouts simple and efficient. We conduct extensive exploration of different strategies for performing multi-step planning on top of the OptFormer model to highlight its potential for use in constructing non-myopic HPO strategies.
Hyperparameter tuning is a fundamental aspect of machine learning research. Setting up the infrastructure for systematic optimization of hyperparameters can take a significant amount of time. Here, we present PyHopper, a black-box optimization platform designed to streamline the hyperparameter tuning workflow of machine learning researchers. PyHopper's goal is to integrate with existing code with minimal effort and run the optimization process with minimal necessary manual oversight. With simplicity as the primary theme, PyHopper is powered by a single robust Markov-chain Monte-Carlo optimization algorithm that scales to millions of dimensions. Compared to existing tuning packages, focusing on a single algorithm frees the user from having to decide between several algorithms and makes PyHopper easily customizable. PyHopper is publicly available under the Apache-2.0 license at https://github.com/PyHopper/PyHopper.
Bayesian Optimization (BO) is a common solution to search optimal hyperparameters based on sample observations of a machine learning model. Existing BO algorithms could converge slowly even collapse when the potential observation noise misdirects the optimization. In this paper, we propose a novel BO algorithm called Neighbor Regularized Bayesian Optimization (NRBO) to solve the problem. We first propose a neighbor-based regularization to smooth each sample observation, which could reduce the observation noise efficiently without any extra training cost. Since the neighbor regularization highly depends on the sample density of a neighbor area, we further design a density-based acquisition function to adjust the acquisition reward and obtain more stable statistics. In addition, we design a adjustment mechanism to ensure the framework maintains a reasonable regularization strength and density reward conditioned on remaining computation resources. We conduct experiments on the bayesmark benchmark and important computer vision benchmarks such as ImageNet and COCO. Extensive experiments demonstrate the effectiveness of NRBO and it consistently outperforms other state-of-the-art methods.
Neural networks are configured by choosing an architecture and hyperparameter values; doing so often involves expert intuition and hand-tuning to find a configuration that extrapolates well without overfitting. This paper considers automatic methods for configuring a neural network that extrapolates in time for the domain of visible and near-infrared (VNIR) spectroscopy. In particular, we study the effect of (a) selecting samples for validating configurations and (b) using ensembles. Most of the time, models are built of the past to predict the future. To encourage the neural network model to extrapolate, we consider validating model configurations on samples that are shifted in time similar to the test set. We experiment with three validation set choices: (1) a random sample of 1/3 of non-test data (the technique used in previous work), (2) using the latest 1/3 (sorted by time), and (3) using a semantically meaningful subset of the data. Hyperparameter optimization relies on the validation set to estimate test-set error, but neural network variance obfuscates the true error value. Ensemble averaging - computing the average across many neural networks - can reduce the variance of prediction errors. To test these methods, we do a comprehensive study of a held-out 2018 harvest season of mango fruit given VNIR spectra from 3 prior years. We find that ensembling improves the state-of-the-art model's variance and accuracy. Furthermore, hyperparameter optimization experiments - with and without ensemble averaging and with each validation set choice - show that when ensembling is combined with using the latest 1/3 of samples as the validation set, a neural network configuration is found automatically that is on par with the state-of-the-art.
Accumulation of corporate data in the cloud has attracted more enterprise applications to the cloud creating data gravity. As a consequence, network traffic has become more cloud centric. This increase in cloud centric traffic poses new challenges in designing learning systems for streaming data due to class imbalance. The number of classes plays a vital role in the accuracy of the classifiers built from the data streams. In this paper, we present a vector quantization-based sampling method, which substantially reduces the class imbalance in data streams. We demonstrate its effectiveness by conducting experiments on network traffic and anomaly dataset with commonly used ML model building methods; Multilayered Perceptron on TensorFlow backend, Support Vector Machines, K-Nearest Neighbour, and Random Forests. We built models using parallel processing, batch processing, and randomly selecting samples. We show that the accuracy of classification models improves when the data streams are pre-processed with our method. We used out of the box hyper-parameters of these classifiers and auto sklearn for hyperparameter optimization.
(1) Background: The success of physiotherapy depends on the regular and correct performance of movement exercises. A system that automatically evaluates these could support the therapy. Previous approaches in this area rarely rely on Deep Learning methods and do not yet fully use their potential. (2) Methods: Using a measurement system consisting of 17 IMUs, a dataset of four Functional Movement Screening (FMS) exercises is recorded. Exercise execution is evaluated by physiotherapists using the FMS criteria. This dataset is used to train a neural network that assigns the correct FMS score to an exercise repetition. We use an architecture consisting of CNN, LSTM and Dense layers. Based on this framework, we apply various methods to optimize the performance of the network. For the optimization, we perform a extensive hyperparameter optimization. In addition, we are comparing different CNN structures that have been specifically adapted for use with IMU data. Finally, the developed network is trained with the data of different FMS exercises and the performance is compared. (3) Results: The evaluation shows that the presented approach achieves a convincing performance in the classification of unknown repetitions of already known subjects. However, the trained network is yet unable to achieve consistent performance on the data of a previously unknown subjects. Additionally, it can be seen that the performance of the network differs significantly depending on the exercise it is trained for.
Sensitive datasets are often underutilized in research and industry due to privacy concerns, limiting the potential of valuable data-driven insights. Synthetic data generation presents a promising solution to address this challenge by balancing privacy protection with data utility. This paper introduces a new approach to mitigate privacy risks associated with outlier observations in synthetic datasets: the Local Resampler (LR). The LR leverages the $k$-nearest neighbors algorithm to generate synthetic data while minimizing disclosure risks by underrepresenting outliers, even when they are not detectable in marginal distributions. Theoretical and empirical analyses demonstrate that the LR effectively mitigates outlier-driven disclosure risks, and accurately replicates multimodal, skewed, and non-convex support distributions. The semiparametric nature of the LR ensures a low computational burden and works efficiently even with small samples. By parameterizing the balance between privacy risks and data utility, this approach promotes broader access to sensitive datasets for research.
Reliable and robust user identification and authentication are important and often necessary requirements for many digital services. It becomes paramount in social virtual reality (VR) to ensure trust, specifically in digital encounters with lifelike realistic-looking avatars as faithful replications of real persons. Recent research has shown that the movements of users in extended reality (XR) systems carry user-specific information and can thus be used to verify their identities. This article compares three different potential encodings of the motion data from head and hands (scene-relative, body-relative, and body-relative velocities), and the performances of five different machine learning architectures (random forest, multi-layer perceptron, fully recurrent neural network, long-short term memory, gated recurrent unit). We use the publicly available dataset "Talking with Hands" and publish all code to allow reproducibility and to provide baselines for future work. After hyperparameter optimization, the combination of a long-short term memory architecture and body-relative data outperformed competing combinations: the model correctly identifies any of the 34 subjects with an accuracy of 100% within 150 seconds. Altogether, our approach provides an effective foundation for behaviometric-based identification and authentication to guide researchers and practitioners. Data and code are published under https://go.uniwue.de/58w1r.
Hyperparameter optimization is the process of identifying the appropriate hyperparameter configuration of a given machine learning model with regard to a given learning task. For smaller data sets, an exhaustive search is possible; However, when the data size and model complexity increase, the number of configuration evaluations becomes the main computational bottleneck. A promising paradigm for tackling this type of problem is surrogate-based optimization. The main idea underlying this paradigm considers an incrementally updated model of the relation between the hyperparameter space and the output (target) space; the data for this model are obtained by evaluating the main learning engine, which is, for example, a factorization machine-based model. By learning to approximate the hyperparameter-target relation, the surrogate (machine learning) model can be used to score large amounts of hyperparameter configurations, exploring parts of the configuration space beyond the reach of direct machine learning engine evaluation. Commonly, a surrogate is selected prior to optimization initialization and remains the same during the search. We investigated whether dynamic switching of surrogates during the optimization itself is a sensible idea of practical relevance for selecting the most appropriate factorization machine-based models for large-scale online recommendation. We conducted benchmarks on data sets containing hundreds of millions of instances against established baselines such as Random Forest- and Gaussian process-based surrogates. The results indicate that surrogate switching can offer good performance while considering fewer learning engine evaluations.
Computing the Jacobian of the solution of an optimization problem is a central problem in machine learning, with applications in hyperparameter optimization, meta-learning, optimization as a layer, and dataset distillation, to name a few. Unrolled differentiation is a popular heuristic that approximates the solution using an iterative solver and differentiates it through the computational path. This work provides a non-asymptotic convergence-rate analysis of this approach on quadratic objectives for gradient descent and the Chebyshev method. We show that to ensure convergence of the Jacobian, we can either 1) choose a large learning rate leading to a fast asymptotic convergence but accept that the algorithm may have an arbitrarily long burn-in phase or 2) choose a smaller learning rate leading to an immediate but slower convergence. We refer to this phenomenon as the curse of unrolling. Finally, we discuss open problems relative to this approach, such as deriving a practical update rule for the optimal unrolling strategy and making novel connections with the field of Sobolev orthogonal polynomials.
Gaussian processes (GPs) are Bayesian non-parametric models popular in a variety of applications due to their accuracy and native uncertainty quantification (UQ). Tuning GP hyperparameters is critical to ensure the validity of prediction accuracy and uncertainty; uniquely estimating multiple hyperparameters in, e.g. the Matern kernel can also be a significant challenge. Moreover, training GPs on large-scale datasets is a highly active area of research: traditional maximum likelihood hyperparameter training requires quadratic memory to form the covariance matrix and has cubic training complexity. To address the scalable hyperparameter tuning problem, we present a novel algorithm which estimates the smoothness and length-scale parameters in the Matern kernel in order to improve robustness of the resulting prediction uncertainties. Using novel loss functions similar to those in conformal prediction algorithms in the computational framework provided by the hyperparameter estimation algorithm MuyGPs, we achieve improved UQ over leave-one-out likelihood maximization while maintaining a high degree of scalability as demonstrated in numerical experiments.
For autonomous skill acquisition, robots have to learn about the physical rules governing the 3D world dynamics from their own past experience to predict and reason about plausible future outcomes. To this end, we propose a transformation-based 3D video prediction (T3VIP) approach that explicitly models the 3D motion by decomposing a scene into its object parts and predicting their corresponding rigid transformations. Our model is fully unsupervised, captures the stochastic nature of the real world, and the observational cues in image and point cloud domains constitute its learning signals. To fully leverage all the 2D and 3D observational signals, we equip our model with automatic hyperparameter optimization (HPO) to interpret the best way of learning from them. To the best of our knowledge, our model is the first generative model that provides an RGB-D video prediction of the future for a static camera. Our extensive evaluation with simulated and real-world datasets demonstrates that our formulation leads to interpretable 3D models that predict future depth videos while achieving on-par performance with 2D models on RGB video prediction. Moreover, we demonstrate that our model outperforms 2D baselines on visuomotor control. Videos, code, dataset, and pre-trained models are available at http://t3vip.cs.uni-freiburg.de.
Bilevel optimization (BO) is useful for solving a variety of important machine learning problems including but not limited to hyperparameter optimization, meta-learning, continual learning, and reinforcement learning. Conventional BO methods need to differentiate through the low-level optimization process with implicit differentiation, which requires expensive calculations related to the Hessian matrix. There has been a recent quest for first-order methods for BO, but the methods proposed to date tend to be complicated and impractical for large-scale deep learning applications. In this work, we propose a simple first-order BO algorithm that depends only on first-order gradient information, requires no implicit differentiation, and is practical and efficient for large-scale non-convex functions in deep learning. We provide non-asymptotic convergence analysis of the proposed method to stationary points for non-convex objectives and present empirical results that show its superior practical performance.