What Is Missing in IRM Training and Evaluation? Challenges and Solutions

Invariant risk minimization (IRM) has received increasing attention as a way to acquire environment-agnostic data representations and predictions, and as a principled solution for preventing spurious correlations from being learned and for improving models' out-of-distribution generalization. Yet, recent works have found that the optimality of the originally-proposed IRM optimization (IRM) may be compromised in practice or could be impossible to achieve in some scenarios. Therefore, a series of advanced IRM algorithms have been developed that show practical improvement over IRM. In this work, we revisit these recent IRM advancements, and identify and resolve three practical limitations in IRM training and evaluation. First, we find that the effect of batch size during training has been chronically overlooked in previous studies, leaving room for further improvement. We propose small-batch training and highlight the improvements over a set of large-batch optimization techniques. Second, we find that improper selection of evaluation environments could give a false sense of invariance for IRM. To alleviate this effect, we leverage diversified test-time environments to precisely characterize the invariance of IRM when applied in practice. Third, we revisit (Ahuja et al. (2020))'s proposal to convert IRM into an ensemble game and identify a limitation when a single invariant predictor is desired instead of an ensemble of individual predictors. We propose a new IRM variant to address this limitation based on a novel viewpoint of ensemble IRM games as consensus-constrained bi-level optimization. Lastly, we conduct extensive experiments (covering 7 existing IRM variants and 7 datasets) to justify the practical significance of revisiting IRM training and evaluation in a principled manner.

Practical Synchronization for OTFS

In the existing literature on joint timing and frequency synchronization of orthogonal time frequency space modulation (OTFS), practically infeasible impulse pilot with large peak-to-average power ratio (PAPR) is deployed. Hence, in this paper, we propose a timing offset (TO) and carrier frequency offset (CFO) estimation for OTFS over a linear time-varying (LTV) channel, using a low PAPR pilot structure. The proposed technique utilizes the recently proposed practically feasible pilot structure with a cyclic prefix (PCP). We exploit the periodic properties of PCP in both delay and time domains to find the starting point of each OTFS block. Furthermore, we propose a two-stage CFO estimation technique with over an order of magnitude higher estimation accuracy than the existing estimator using the impulse pilot. In the first stage, a coarse CFO estimate is obtained which is refined in the second stage, through our proposed maximum likelihood (ML) based approach. The proposed ML-based approach deploys the generalized complex exponential basis expansion model (GCE-BEM) to capture the time variations of the channel, absorb them into the pilot and provide an accurate CFO estimate. Since our proposed synchronization technique utilizes the same pilot deployed for channel estimation, it does not require any additional overhead. Finally, we evaluate the performance of our proposed synchronization technique through simulations. We also compare and show the superior performance of our proposed technique to the only other existing joint TO and CFO estimation method in OTFS literature.

Multi-Scale Message Passing Neural PDE Solvers

We propose a novel multi-scale message passing neural network algorithm for learning the solutions of time-dependent PDEs. Our algorithm possesses both temporal and spatial multi-scale resolution features by incorporating multi-scale sequence models and graph gating modules in the encoder and processor, respectively. Benchmark numerical experiments are presented to demonstrate that the proposed algorithm outperforms baselines, particularly on a PDE with a range of spatial and temporal scales.

On the Detection and Quantification of Nonlinearity via Statistics of the Gradients of a Black-Box Model

Detection and identification of nonlinearity is a task of high importance for structural dynamics. Detecting nonlinearity in a structure, which has been designed to operate in its linear region, might indicate the existence of damage. Therefore, it is important, even for safety reasons, to detect when a structure exhibits nonlinear behaviour. In the current work, a method to detect nonlinearity is proposed, based on the distribution of the gradients of a data-driven model, which is fitted on data acquired from the structure of interest. The data-driven model herein is a neural network. The selection of such a type of model was done in order to not allow the user to decide how linear or nonlinear the model shall be, but to let the training algorithm of the neural network shape the level of nonlinearity according to the training data. The neural network is trained to predict the accelerations of the structure for a time-instant using as inputs accelerations of previous time-instants, i.e. one-step-ahead predictions. Afterwards, the gradients of the output of the neural network with respect to its inputs are calculated. Given that the structure is linear, the distribution of the aforementioned gradients should be quite peaked, while in the case of a structure with nonlinearities, the distribution of the gradients shall be more spread and, potentially, multimodal. To test the above assumption, data from an experimental structure are considered. The structure is tested under different scenarios, some of which are linear and some nonlinear. The statistics of the distributions of the gradients for the different scenarios can be used to identify cases where nonlinearity is present. Moreover, via the proposed method one is able to quantify the nonlinearity by observing higher values of standard deviation of the distribution of the gradients for "more nonlinear" scenarios.

Learning Vision-based Robotic Manipulation Tasks Sequentially in Offline Reinforcement Learning Settings

With the rise of deep reinforcement learning (RL) methods, many complex robotic manipulation tasks are being solved. However, harnessing the full power of deep learning requires large datasets. Online-RL does not suit itself readily into this paradigm due to costly and time-taking agent environment interaction. Therefore recently, many offline-RL algorithms have been proposed to learn robotic tasks. But mainly, all such methods focus on a single task or multi-task learning, which requires retraining every time we need to learn a new task. Continuously learning tasks without forgetting previous knowledge combined with the power of offline deep-RL would allow us to scale the number of tasks by keep adding them one-after-another. In this paper, we investigate the effectiveness of regularisation-based methods like synaptic intelligence for sequentially learning image-based robotic manipulation tasks in an offline-RL setup. We evaluate the performance of this combined framework against common challenges of sequential learning: catastrophic forgetting and forward knowledge transfer. We performed experiments with different task combinations to analyze the effect of task ordering. We also investigated the effect of the number of object configurations and density of robot trajectories. We found that learning tasks sequentially helps in the propagation of knowledge from previous tasks, thereby reducing the time required to learn a new task. Regularisation based approaches for continuous learning like the synaptic intelligence method although helps in mitigating catastrophic forgetting but has shown only limited transfer of knowledge from previous tasks.

A Proof that Using Crossover Can Guarantee Exponential Speed-Ups in Evolutionary Multi-Objective Optimisation

Evolutionary algorithms are popular algorithms for multiobjective optimisation (also called Pareto optimisation) as they use a population to store trade-offs between different objectives. Despite their popularity, the theoretical foundation of multiobjective evolutionary optimisation (EMO) is still in its early development. Fundamental questions such as the benefits of the crossover operator are still not fully understood. We provide a theoretical analysis of well-known EMO algorithms GSEMO and NSGA-II to showcase the possible advantages of crossover. We propose a class of problems on which these EMO algorithms using crossover find the Pareto set in expected polynomial time. In sharp contrast, they and many other EMO algorithms without crossover require exponential time to even find a single Pareto-optimal point. This is the first example of an exponential performance gap through the use of crossover for the widely used NSGA-II algorithm.

Differentially-Private Hierarchical Clustering with Provable Approximation Guarantees

Hierarchical Clustering is a popular unsupervised machine learning method with decades of history and numerous applications. We initiate the study of differentially private approximation algorithms for hierarchical clustering under the rigorous framework introduced by (Dasgupta, 2016). We show strong lower bounds for the problem: that any $\epsilon$-DP algorithm must exhibit $O(|V|^2/ \epsilon)$-additive error for an input dataset $V$. Then, we exhibit a polynomial-time approximation algorithm with $O(|V|^{2.5}/ \epsilon)$-additive error, and an exponential-time algorithm that meets the lower bound. To overcome the lower bound, we focus on the stochastic block model, a popular model of graphs, and, with a separation assumption on the blocks, propose a private $1+o(1)$ approximation algorithm which also recovers the blocks exactly. Finally, we perform an empirical study of our algorithms and validate their performance.

Combining Tree-Search, Generative Models, and Nash Bargaining Concepts in Game-Theoretic Reinforcement Learning

Multiagent reinforcement learning (MARL) has benefited significantly from population-based and game-theoretic training regimes. One approach, Policy-Space Response Oracles (PSRO), employs standard reinforcement learning to compute response policies via approximate best responses and combines them via meta-strategy selection. We augment PSRO by adding a novel search procedure with generative sampling of world states, and introduce two new meta-strategy solvers based on the Nash bargaining solution. We evaluate PSRO's ability to compute approximate Nash equilibrium, and its performance in two negotiation games: Colored Trails, and Deal or No Deal. We conduct behavioral studies where human participants negotiate with our agents ($N = 346$). We find that search with generative modeling finds stronger policies during both training time and test time, enables online Bayesian co-player prediction, and can produce agents that achieve comparable social welfare negotiating with humans as humans trading among themselves.

Multiagent Inverse Reinforcement Learning via Theory of Mind Reasoning

We approach the problem of understanding how people interact with each other in collaborative settings, especially when individuals know little about their teammates, via Multiagent Inverse Reinforcement Learning (MIRL), where the goal is to infer the reward functions guiding the behavior of each individual given trajectories of a team's behavior during some task. Unlike current MIRL approaches, we do not assume that team members know each other's goals a priori; rather, that they collaborate by adapting to the goals of others perceived by observing their behavior, all while jointly performing a task. To address this problem, we propose a novel approach to MIRL via Theory of Mind (MIRL-ToM). For each agent, we first use ToM reasoning to estimate a posterior distribution over baseline reward profiles given their demonstrated behavior. We then perform MIRL via decentralized equilibrium by employing single-agent Maximum Entropy IRL to infer a reward function for each agent, where we simulate the behavior of other teammates according to the time-varying distribution over profiles. We evaluate our approach in a simulated 2-player search-and-rescue operation where the goal of the agents, playing different roles, is to search for and evacuate victims in the environment. Our results show that the choice of baseline profiles is paramount to the recovery of the ground-truth rewards, and that MIRL-ToM is able to recover the rewards used by agents interacting both with known and unknown teammates.

Online Video Streaming Super-Resolution with Adaptive Look-Up Table Fusion

This paper focuses on Super-resolution for online video streaming data. Applying existing super-resolution methods to video streaming data is non-trivial for two reasons. First, to support application with constant interactions, video streaming has a high requirement for latency that most existing methods are less applicable, especially on low-end devices. Second, existing video streaming protocols (e.g., WebRTC) dynamically adapt the video quality to the network condition, thus video streaming in the wild varies greatly under different network bandwidths, which leads to diverse and dynamic degradations. To tackle the above two challenges, we proposed a novel video super-resolution method for online video streaming. First, we incorporate Look-Up Table (LUT) to lightweight convolution modules to achieve real-time latency. Second, for variant degradations, we propose a pixel-level LUT fusion strategy, where a set of LUT bases are built upon state-of-the-art SR networks pre-trained on different degraded data, and those LUT bases are combined with extracted weights from lightweight convolution modules to adaptively handle dynamic degradations. Extensive experiments are conducted on a newly proposed online video streaming dataset named LDV-WebRTC. All the results show that our method significantly outperforms existing LUT-based methods and offers competitive SR performance with faster speed compared to efficient CNN-based methods. Accelerated with our parallel LUT inference, our proposed method can even support online 720P video SR around 100 FPS.