United States

We study model-based reinforcement learning (RL) for episodic Markov decision processes (MDP) whose transition probability is parametrized by an unknown transition core with features of state and action. 
Despite much recent progress in analyzing algorithms in the linear MDP setting, the understanding of more general transition models is very restrictive. 
In this paper, we establish a provably efficient RL algorithm for the MDP 
whose state transition is given by a multinomial logistic model. 
To balance the exploration-exploitation trade-off, we propose an upper confidence bound-based algorithm.
We show that our proposed algorithm achieves $\tilde{\mathcal{O}}(d \sqrt{H^3 T})$ regret bound where $d$ is the dimension of the transition core, $H$ is the horizon, and $T$ is the total number of steps. 
To the best of our knowledge, this is the first model-based RL algorithm with multinomial logistic function approximation with provable guarantees.
We also comprehensively evaluate our proposed algorithm numerically and show that it consistently outperforms the existing methods, hence achieving both provable efficiency and practical superior performance.

AAAI 2023

Model-Based Reinforcement Learning with Multinomial Logistic Function Approximation

multinomial logistic model

regret analysis

reinforcement learning

The purpose of the AAAI conference series is to promote research in Artificial Intelligence (AI) and foster scientific exchange between researchers, practitioners, scientists, students, and engineers across the entirety of AI and its affiliated disciplines. AAAI-23 is the Thirty-Seventh AAAI Conference on Artificial Intelligence. The theme of this conference is to create collaborative bridges within and beyond AI. Like previous AAAI conferences, AAAI-23 will feature technical paper presentations, special tracks, invited speakers, workshops, tutorials, poster sessions, senior member presentations, competitions, and exhibit programs, and two new activities: a Bridge Program and a Lab Program. Many of these activities are tailored to the theme of bridges and all are selected according to the highest standards, with additional programs for students and young researchers. 
AAAI is providing you with a conference planner, which you can use to help organize your itinerary of activities. This includes talks to attend in person, talks to attend remotely, breaks with colleagues and your site seeing activities. To access this conference planner, please go to [https://aaai-2023.takemobi.io](https://aaai-2023.takemobi.io).

In order to access this site, you need to register. If you haven't already, please register [here](https://aaai.org/Conferences/AAAI-23/registration/).


The purpose of the AAAI conference series is to promote research in Artificial Intelligence (AI) and foster scientific exchange between researchers, practitioners, scientists, students, and engineers across the entirety of AI and its affiliated disciplines.

poster

Label distribution covers a certain number of labels, representing the degree to which each label describes an instance. The learning process on the instances labeled by label distributions is called Label Distribution Learning (LDL). Although LDL has been applied successfully to many practical applications, one problem with existing LDL methods is that they are limited to data with balanced label information. However, annotation information in real-world data often exhibits imbalanced distributions, which significantly degrades the performance of existing methods. In this paper, we investigate the Imbalanced Label Distribution Learning (ILDL) problem. To handle this challenging problem, we delve into the characteristics of ILDL and empirically find that the representation distribution shift is the underlying reason for the performance degradation of existing methods. Inspired by this finding, we present a novel method named Representation Distribution Alignment (RDA). RDA aligns the distributions of feature representations and label representations to alleviate the impact of the distribution gap between the training set and the test set caused by the imbalance issue. Extensive experiments verify the superior performance of RDA. Our work fills the gap in benchmarks and techniques for practical ILDL problems.

Imbalanced Label Distribution Learning

Model inversion (MI) attacks have raised increasing concerns about privacy, which can reconstruct training data from public models. Indeed, MI attacks can be formalized as an optimization problem that seeks private data in a certain space. Recent MI attacks leverage a generative adversarial network (GAN) as an image prior to narrow the search space, and can successfully reconstruct even the high-dimensional data (e.g., face images). However, these generative MI attacks do not fully exploit the potential capabilities of the target model, still leading to a vague and coupled search space, i.e., different classes of images are coupled in the search space. Besides, the widely used cross-entropy loss in these attacks suffers from gradient vanishing. To address these problems, we propose Pseudo Label-Guided MI (PLG-MI) attack via conditional GAN (cGAN). At first, a top-\emph{n} selection strategy is proposed to provide pseudo-labels for public data, and use pseudo-labels to guide the training of the cGAN. In this way, the search space is decoupled for different classes of images. Then a max-margin loss is introduced to improve the search process on the subspace of a target class. Extensive experiments demonstrate that our PLG-MI attack significantly improves the attack success rate and visual quality for various datasets and models, notably, $2\sim3 \times$ better than state-of-the-art attacks under large distributional shifts.

Pseudo Label-Guided Model Inversion Attack via Conditional Generative Adversarial Network

Evaluation of generative models is mostly based on the comparison between the estimated distribution and the ground truth distribution in a certain feature space. To embed samples into informative features, previous works often use convolutional neural networks optimized for classification, which is criticized by recent studies. Therefore, various feature spaces have been explored to discover alternatives. Among them, a surprising approach is to use a randomly initialized neural network for feature embedding. However, the fundamental basis to employ the random features has not been sufficiently justified. In this paper, we rigorously investigate the feature space of models with random weights in comparison to that of trained models. Furthermore, we provide an empirical evidence to choose networks for random features to obtain consistent and reliable results. Our results indicate that the features from random networks can evaluate generative models well similarly to those from trained networks, and furthermore, the two types of features can be used together in a complementary way.

Demystifying Randomly Initialized Networks for Evaluating Generative Models

This paper proposes an unsupervised multi-exposure image fusion (MEF) method via contrastive learning, termed as MEF-CL. It breaks exposure limits and performance bottleneck faced by existing methods. MEF-CL firstly designs similarity constraints to preserve contents in source images. It eliminates the need for ground truth (actually not exist and created artificially) and thus avoids negative impacts of inappropriate ground truth on performance and generalization. Moreover, we explore a latent feature space and apply contrastive learning in this space to guide fused image to approximate normal-light samples and stay away from inappropriately exposed ones. In this way, characteristics of fused images (e.g., illumination, colors) can be further improved without being subject to source images. Therefore, MEF-CL is applicable to image pairs of any multiple exposures rather than a pair of under-exposed and over-exposed images mandated by existing methods. By alleviating dependence on source images, MEF-CL shows better generalization for various scenes. Consequently, our results exhibit appropriate illumination, detailed textures, and saturated colors. Qualitative, quantitative, and ablation experiments validate the superiority and generalization of MEF-CL. Our code is publicly available at https://github.com/hanna-xu/MEF-CL.

Unsupervised Multi-exposure Image Fusion Breaking Exposure Limits via Contrastive Learning

We study various models for the one-dimensional multi-stage facility location problems with transient agents, where a transient agent arrives in some stage and stays for a number of consecutive stages. In the problems, we need to serve each agent in one of their stages by determining the location of the facility at each stage. In the first model, we assume there is no cost for moving the facility across the stages. We focus on optimal algorithms to minimize both the social cost objective, defined as the total distance of all agents to the facility over all stages, and the maximum cost objective, defined as the max distance of any agent to the facility over all stages. For each objective, we give a slice-wise polynomial (XP) algorithm (i.e., solvable in m^f(k) for some fixed parameter k and computable function f, where m is the input size) and show that there is a polynomial-time algorithm when a natural first-come-first-serve (FCFS) order of agent serving is enforced. We then consider the mechanism design problem, where the agents' locations and arrival stages are private, and design a group strategy-proof mechanism that achieves good approximation ratios for both objectives and settings with and without FCFS ordering. In the second model, we consider the facility's moving cost between adjacent stages under the social cost objective, which accounts for the total moving distance of the facility. Correspondingly, we design XP (and polynomial time) algorithms and a group strategy-proof mechanism for settings with or without the FCFS ordering.  

Multi-Stage Facility Location Problems with Transient Agents

Amodal instance segmentation aims to infer the amodal mask, including both the visible part and occluded part of each object instance. Predicting the occluded parts is challenging. Existing methods often produce incomplete amodal boxes and amodal masks, probably due to lacking visual evidences to expand the boxes and masks. To this end, we propose a prior-guided expansion framework, which builds on a two-stage segmentation model (i.e., Mask R-CNN) and performs box-level (resp., pixel-level) expansion for amodal box (resp., mask) prediction, by retrieving regression (resp., flow) transformations from a memory bank of expansion prior. We conduct extensive experiments on KINS, D2SA, and COCOA cls datasets, which show the effectiveness of our method.

Amodal Instance Segmentation via Prior-guided Expansion

3D object detection in autonomous driving aims to reason “what” and “where” the objects of interest present in a 3D world. Following the conventional wisdom of previous 2D object detection, existing methods often adopt the canonical Cartesian coordinate system with perpendicular axis. However, we conjugate that this does not fit the nature of the ego car’s perspective, as each onboard camera perceives the world in shape of wedge intrinsic to the imaging geometry with radical (non perpendicular) axis. Hence, in this paper we advocate the exploitation of the Polar coordinate system and propose a new Polar Transformer (PolarFormer) for more accurate 3D object detection in the bird’s-eye-view (BEV) taking as input only multi-camera 2D images. Specifically, we design a cross-attention based Polar detection head without restriction to the shape of input structure to deal with irregular Polar grids. For tackling the unconstrained object scale variations along Polar’s distance dimension, we further introduce a multi-scale Polar representation learning strategy. As a result, our model can make best use of the Polar representation rasterized via attending to the corresponding image observation in a sequence-to-sequence fashion subject to the geometric constraints. Thorough experiments on the nuScenes dataset demonstrate that our PolarFormer outperforms significantly state-of-the-art 3D object detection alternatives.

PolarFormer: Multi-camera 3D Object Detection with Polar Transformer

We study the problem of training and certifying adversarially robust quantized neural networks (QNNs). Quantization is a technique for making neural networks more efficient by running them using low-bit integer arithmetic and is therefore commonly adopted in industry. Recent work has shown that floating-point neural networks that have been verified to be robust can become vulnerable to adversarial attacks after quantization, and certification of the quantized representation is necessary to guarantee robustness.
In this work, we present quantization-aware interval bound propagation (QA-IBP), a novel method for training robust QNNs.
Inspired by advances in robust learning of non-quantized networks, our training algorithm computes the gradient of an abstract representation of the actual network. Unlike existing approaches, our method can handle the discrete semantics of QNNs. 
Based on QA-IBP, we also develop a complete verification procedure for verifying the adversarial robustness of QNNs, which is guaranteed to terminate and produce a correct answer. Compared to existing approaches, the key advantage of our verification procedure is that it runs entirely on GPU or other accelerator devices. 
We demonstrate experimentally that our approach significantly outperforms existing methods and establish the new state-of-the-art for training and certifying the robustness of QNNs.

Premium content

Model-Based Reinforcement Learning with Multinomial Logistic Function Approximation

Downloads

Next from AAAI 2023

Imbalanced Label Distribution Learning

Similar lecture

Cooperative Multi-Agent Learning in a Complex World: Challenges and Solutions