Daily TMLR digest for Feb 24, 2023

0 views
Skip to first unread message

TMLR

unread,
Feb 23, 2023, 7:00:11 PM2/23/23
to tmlr-anno...@googlegroups.com


Accepted papers
===============


Title: A Flexible Nadaraya-Watson Head Can Offer Explainable and Calibrated Classification

Authors: Alan Q. Wang, Mert R. Sabuncu

Abstract: In this paper, we empirically analyze a simple, non-learnable, and nonparametric Nadaraya-Watson (NW) prediction head that can be used with any neural network architecture. In the NW head, the prediction is a weighted average of labels from a support set. The weights are computed from distances between the query feature and support features. This is in contrast to the dominant approach of using a learnable classification head (e.g., a fully-connected layer) on the features, which can be challenging to interpret and can yield poorly calibrated predictions. Our empirical results on an array of computer vision tasks demonstrate that the NW head can yield better calibration with comparable accuracy compared to its parametric counterpart, particularly in data-limited settings. To further increase inference-time efficiency, we propose a simple approach that involves a clustering step run on the training set to create a relatively small distilled support set. Furthermore, we explore two means of interpretability/explainability that fall naturally from the NW head. The first is the label weights, and the second is our novel concept of the ``support influence function,'' which is an easy-to-compute metric that quantifies the influence of a support element on the prediction for a given query. As we demonstrate in our experiments, the influence function can allow the user to debug a trained model. We believe that the NW head is a flexible, interpretable, and highly useful building block that can be used in a range of applications.

URL: https://openreview.net/forum?id=iEq6lhG4O3

---

Title: Diffusion-based Time Series Imputation and Forecasting with Structured State Space Models

Authors: Juan Lopez Alcaraz, Nils Strodthoff

Abstract: The imputation of missing values represents a significant obstacle for many real-world data analysis pipelines. Here, we focus on time series data and put forward SSSD, an imputation model that relies on two emerging technologies, (conditional) diffusion models as state-ofthe-art generative models and structured state space models as internal model architecture, which are particularly suited to capture long-term dependencies in time series data. We demonstrate that SSSD matches or even exceeds state-of-the-art probabilistic imputation and forecasting performance on a broad range of data sets and different missingness scenarios, including the challenging blackout-missing scenarios, where prior approaches failed to provide meaningful results.

URL: https://openreview.net/forum?id=hHiIbk7ApW

---


New submissions
===============


Title: Comparative Generalization Bounds for Deep Neural Networks

Abstract: In this work, we investigate the generalization capabilities of deep neural networks. We introduce a measure of the effective depth of neural networks, defined as the first layer at which sample embeddings are separable using the nearest-class center classifier. Our empirical results demonstrate that, in standard classification settings, neural networks trained using Stochastic Gradient Descent tend to have small effective depths. We also explore the relationship between effective depth, the complexity of the training dataset, and generalization. For instance, we find that the effective depth of a trained neural network increases as the number of random labels in the data increases. Additionally, we derive a generalization bound by comparing the effective depth of a network with the minimal depth required to fit the same dataset with partially corrupted labels. This bound provides non-vacuous predictions of test performance and is found to be independent of the actual depth of the network in our experiments.

URL: https://openreview.net/forum?id=162TqkUNPO

---

Title: Evolving Pareto-Optimal Actor-Critic Algorithms for Generalizability and Stability

Abstract: Generalizability and stability are two key objectives for operating reinforcement learning (RL) agents in the real world. Designing RL algorithms that optimize these objectives can be a costly and painstaking process. This paper presents MetaPG, an evolutionary method for automated design of actor-critic loss functions. MetaPG explicitly optimizes for generalizability and performance, and implicitly optimizes the stability of both metrics. We initialize our loss function population with Soft Actor-Critic (SAC) and perform multi-objective optimization using fitness metrics encoding single-task performance, zero-shot generalizability to unseen environment configurations, and stability across independent runs with different random seeds. On a set of continuous control tasks from the Real-World RL Benchmark Suite, we find that our method, using a single environment during evolution, evolves algorithms that improve upon SAC's performance and generalizability by 4% and 20%, respectively, and reduce instability up to 67%. Then, we scale up to more complex environments from the Brax physics simulator and replicate generalizability tests encountered in practical settings, such as different friction coefficients. MetaPG evolves algorithms that can obtain 10% better generalizability without loss of performance within the same meta-training environment and obtain similar results to SAC when doing cross-domain evaluations in other Brax environments. The evolution results are interpretable; by analyzing the structure of the best algorithms we identify elements that help optimizing certain objectives, such as regularization terms for the critic loss.

URL: https://openreview.net/forum?id=zTklDjybrK

---
Reply all
Reply to author
Forward
0 new messages