United States

Incorporating sequence-to-sequence models into history-based Reinforcement Learning (RL) provides a general way to extend RL to partially-observable tasks. This method compresses history spaces according to the correlations between historical observations and the rewards. However, they do not adjust for the confounding correlations caused by data sampling and assign high beliefs to uninformative historical observations, leading to limited compression of history spaces. Counterfactual Inference (CI), which estimates causal effects by single-variable intervention, is a promising way to adjust for confounding. However, it is computationally infeasible to directly apply the single-variable intervention to a huge number of historical observations. This paper proposes to perform CI on observation sub-spaces instead of single observations and develop a coarse-to-fine CI algorithm, called Tree-based History Counterfactual Inference (T-HCI), to reduce the number of interventions exponentially. We show that T-HCI is computationally feasible in practice and brings significant sample efficiency gains in various challenging partially-observable tasks, including Maze, BabyAI, and robot manipulation tasks.

AAAI 2023

Fast Counterfactual Inference for History-Based Reinforcement Learning

sequential decision making

reinforcement learning algorithms

causal 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

Voxel grid representation of 3D scene properties has been widely used to improve the training or rendering speed of the Neural Radiance Fields (NeRF) while at the same time achieving high synthesis quality. However, these methods accelerate the original NeRF at the expense of extra storage demand, which hinders their applications in many scenarios. To solve this limitation, we present TinyNeRF, a three-stage pipeline: frequency domain transformation, pruning and quantization that work together to reduce the storage demand of the voxel grids with little to no effects on their speed and synthesis quality. Based on the prior knowledge of visual signals sparsity in the frequency domain, we convert the original voxel grids in the frequency domain via block-wise discrete cosine transformation (DCT). Next, we apply pruning and quantization to enforce the DCT coefficients to be sparse and low-bit. Our method can be optimized from scratch in an end-to-end manner, and can typically compress the original models by 2 orders of magnitude with minimal sacrifice on speed and synthesis quality.

TinyNeRF: Towards $100\times$ Compression of Voxel Radiance Fields

3D object detection received increasing attention in autonomous driving recently. Objects in 3D scenes are distributed with diverse orientations. Ordinary detectors do not explicitly model the variations of rotation and reflection transformations. Consequently, large networks and extensive data augmentation are required for robust detection. Recent equivariant networks explicitly model the transformation variations by applying shared networks on multiple transformed point clouds, showing great potential in object geometry modeling. However, it is difficult to apply such networks to 3D object detection in autonomous driving due to its large computation cost and slow reasoning speed. In this work, we present TED, an efficient Transformation-Equivariant 3D Detector to overcome the computation cost and speed issues. TED first applies a sparse convolution backbone to extract multi-channel transformation-equivariant voxel features; and then aligns and aggregates these equivariant features into lightweight and compact representations for high-performance 3D object detection. On the highly competitive KITTI 3D car detection leaderboard, TED ranked 1st among all submissions with competitive efficiency.

Transformation-Equivariant 3D Object Detection for Autonomous Driving

Positive-unlabeled learning is an essential problem in many real-world applications with only labeled positive and unlabeled data, especially when the negative samples are difficult to identify. Most existing positive-unlabeled learning methods will inevitably overfit the positive class to some extent due to the existence of unidentified positive samples. This paper first analyzes the overfitting problem and proposes to bound the generalization errors via Wasserstein distances. Based on that, we develop a simple yet effective positive-unlabeled learning method, GradPU, which consists of two key ingredients: A gradient-based regularizer that penalizes the gradient norms in the interpolated data region, which improves the generalization of positive class; An unnormalized upweighting mechanism that assigns larger weights to those positive samples that are hard, not-well-fitted and less frequently labeled. It enforces the training error of each positive sample to be small and increases the robustness to the labeling bias. We evaluate our proposed GradPU on three datasets: MNIST, FashionMNIST, and CIFAR10. The results demonstrate that GradPU achieves state-of-the-art performance on both unbiased and biased positive labeling scenarios.

GradPU: Positive-Unlabeled Learning via Gradient Penalty and Positive Upweighting

We study convex Constrained Markov Decision Processes (CMDPs) in which the objective is concave and the constraints are convex in the state-action occupancy measure. We propose a policy-based primal-dual algorithm that updates the primal variable via policy gradient ascent and updates the dual variable via projected sub-gradient descent. Despite the loss of additivity structure and the nonconvex nature, we establish the global convergence of the proposed algorithm by leveraging a hidden convexity in the problem, and prove the $O(T^{-1/3})$ convergence rate in terms of both optimality gap and constraint violation. When the objective is strongly concave in the occupancy measure, we prove an improved convergence rate of $O(T^{-1/2})$. By introducing a pessimistic term to the constraint, we further show that a zero constraint violation can be achieved while preserving the same convergence rate for the optimality gap. This work is the first one in the literature that establishes non-asymptotic convergence guarantees for policy-based primal-dual methods for solving infinite-horizon discounted convex CMDPs.

Policy-based Primal-Dual Methods for Convex Constrained Markov Decision Processes

In this work, we propose a real-time monocular 3D video reconstruction approach named Flora for reconstructing delicate and complete 3D scenes from RGB video sequences in an end-to-end manner. Specifically, we introduce a novel method with two main contributions. Firstly, the proposed feature aggregation module retains both color and reliability in a dual-frequency form. Secondly, the loss compensation module solves missing structure by correcting losses for falsely pruned voxels. The dual-frequency feature aggregation module enhances reconstruction quality in both precision and recall, and the loss compensation module benefits the recall. Notably, both proposed contributions achieve great results with negligible inferencing overhead. Our state-of-the-art experimental results on real-world datasets demonstrate Flora's leading performance in both effectiveness and efficiency. The code is available at https://github.com/NoOneUST/Flora.

Flora: dual-Frequency LOss-compensated ReAl-time monocular 3D video reconstruction

We study the online Traveling Salesman Problem (TSP) on the line augmented with machine-learned predictions. In the classical problem, there is a stream of requests released over time along the real line. The goal is to minimize the makespan of the algorithm. We distinguish between the open variant and the closed one, in which we additionally require the algorithm to return to the origin after serving all requests. The state of the art is a 1.64-competitive algorithm and a 2.04-competitive algorithm for the closed and open variants, respectively. In both cases, a tight lower bound is known.

In both variants, our primary prediction model involves predicted positions of the requests. We introduce algorithms that (i) obtain a tight 1.5 competitive ratio for the closed variant and a 1.66 competitive ratio for the open variant in the case of perfect predictions, (ii) are robust against unbounded prediction error, and (iii) are smooth, i.e., their performance degrades gracefully as the prediction error increases.

Moreover, we further investigate the learning-augmented setting in the open variant by additionally considering a prediction for the last request served by the optimal offline algorithm. Our algorithm for this enhanced setting obtains a 1.33 competitive ratio with perfect predictions while also being smooth and robust, beating the lower bound of 1.44 we show for our original prediction setting for the open variant. Also, we provide a lower bound of 1.25 for this enhanced setting.

Learning-Augmented Algorithms for Online TSP on the Line

Monaural speech separation aims to separate concurrent speakers from a single-microphone mixture recording. Inspired by the effect of pitch priming in auditory scene analysis (ASA) mechanisms, a novel pitch-guided speech separation framework is proposed in this work. The prominent advantage of this framework is that both the permutation problem and the unknown speaker number problem existing in general models can be avoided by using pitch contours as the primary means to guide the target speaker. In addition, adversarial training is applied, instead of a traditional time-frequency mask, to improve the perceptual quality of separated speech. Specifically, the proposed framework can be divided into two phases: pitch extraction and speech separation. The former aims to extract pitch contour candidates for each speaker from the mixture, modeling the bottom-up process in ASA mechanisms. Any pitch contour can be selected as the condition in the second phase to separate the corresponding speaker, where a conditional generative adversarial network (CGAN) is applied. The second phase models the effect of pitch priming in ASA. Experiments on the WSJ0-2mix corpus reveal that the proposed approaches can achieve higher pitch extraction accuracy and better separation performance, compared to the baseline models, and have the potential to be applied to SOTA architectures.

PGSS: Pitch-Guided Speech Separation

Previous works on emotion recognition in conversation (ERC) follow a two-step paradigm, which can be summarized as first producing context-independent features via fine-tuning pretrained language models (PLMs) and then analyzing contextual information and dialogue structure information among the extracted features. However, we discover that this paradigm has several limitations. Accordingly, we propose a novel paradigm, i.e., exploring contextual information and dialogue structure information in the fine-tuning step, and adapting the PLM to the ERC task in terms of input text, classification structure, and training strategy. Furthermore, we develop our model BERT-ERC according to the proposed paradigm, which improves ERC performance in three aspects, namely suggestive text, fine-grained classification module, and two-stage training. Compared to existing methods, BERT-ERC achieves substantial improvement on four datasets, indicating its effectiveness and generalization capability. Besides, we also set up the limited resources scenario and the online prediction scenario to approximate real-world scenarios. Extensive experiments demonstrate that the proposed paradigm significantly outperforms the previous one and can be adapted to various scenes.

BERT-ERC: Fine-tuning BERT is Enough for Emotion Recognition in Conversation

Sophisticated data augmentations can greatly boost the efficiency of Graph Contrastive Learning (GCL). However, there are works investigating the feature augmentation for unsupervised contrastive learning, although (class-conditional) feature augmentation has been proven useful in supervised learning. Thus, we present a novel/efficient spectral feature argumentation for contrastive learning on graphs (and images). To this end, from two data views, respectively, we estimate a low-rank approximation per feature map and subtract that approximation from the map to obtain its complement. This is achieved by the proposed herein so-called incomplete power iteration, a non-standard power iteration regime which enjoys two valuable byproducts (under the low number of iterations only): (i) relieve the unbalancing spectrum of the feature map, and (ii) injecting the noise into the singular values of the feature map. For two views, we align these complement feature maps as the spectrum-balanced alignment step focuses on less dominant singular values of spectra of both views, whereas the spectral augmentation of singular values does not affect the spectral angle alignment (singular vectors are left intact). We derive the analytical form for the incomplete power iteration to capture its spectrum-balancing effect and the variance of singular values implicitly augmented by the noise. We also show that this augmentation leads to a lower generalization bound. Experiments on graph and image datasets demonstrate the proposed spectral feature augmentation outperforms baselines, and is complementary with other graph augmentation approaches and compatible with various contrastive losses.

Spectral Feature Augmentation for Graph Contrastive Learning and Beyond

Pruning has been an effective solution to reduce the number of computations and the memory requirement in deep learning.
The pruning unit plays an important role in exploiting the GPU resources efficiently. 
The filter is proposed as a simple pruning unit of structured pruning.
However, since the filter is quite large as pruning unit, the accuracy drop is considerable with a high pruning ratio.
GPU rearranges the weight and input tensors into tiles (blocks) for efficient computation. 
To fully utilize GPU resources, this tile structure should be considered, which is the goal of block pruning. 
However, previous block pruning prunes both row vectors and column vectors. 
Pruning of row vectors in a tile corresponds to filter pruning, and it also interferes with column-wise block pruning of the following layer.
In contrast, column vectors are much smaller than row vectors and can achieve lower accuracy drop.
Additionally, if the pruning ratio for each tile is different,
GPU utilization can be limited by imbalanced workloads by irregular-sized blocks.
The same pruning ratio for the weight tiles processed in parallel enables the actual inference process to fully utilize the resources without idle time.
This paper proposes balanced column-wise block pruning, named BCBP, to satisfy two conditions: the column-wise minimal size of the pruning unit and balanced workloads. 
We demonstrate that BCBP is superior to previous pruning methods through comprehensive experiments.

Fast Counterfactual Inference for History-Based Reinforcement Learning

Downloads

Next from AAAI 2023

TinyNeRF: Towards $100\times$ Compression of Voxel Radiance Fields

Similar lecture

Safe Reinforcement Learning via Shielding under Partial Observability