United States

 Lottery tickets (LTs) is able to discover accurate and sparse subnetworks that could be trained in isolation to match the performance of dense networks. Ensemble, in parallel, is one of the oldest time-proven tricks in machine learning to improve performance by combining the output of multiple independent models. However, the benefits of ensemble in the context of LTs will be diluted since ensemble does not directly lead to stronger sparse subnetworks, but leverages their predictions for a better decision. In this work, we first observe that directly averaging the weights of the adjacent learned subnetworks significantly boosts the performance of LTs. Encouraged by this observation, we further propose an alternative way to perform an ``ensemble&#39;&#39; over the subnetworks identified by iterative magnitude pruning via a simple interpolating strategy. We call our method &quot;Lottery Pools&quot;. In contrast to the naive ensemble which brings no performance gains to each single subnetwork, Lottery Pools yields much stronger sparse subnetworks than the original LTs without requiring any extra training or inference cost. Across various modern architectures on CIFAR-10/100 and ImageNet, we show that our method achieves significant performance gains in both, in-distribution and out-of-distribution scenarios. Impressively, evaluated with VGG-16 and ResNet-18, the produced sparse subnetworks outperform the original LTs by up to 1.88% on CIFAR-100 and 2.36\% on CIFAR-100-C; the resulting dense network surpasses the pre-trained dense-model up to 2.22% on CIFAR-100 and 2.38% on CIFAR-100-C.

AAAI 2023

Lottery Pools: Winning More by Interpolating Tickets without Increasing Training or Inference Cost

weight interpolating

lottery tickets hypothesis

model efficiency

computer vision

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.

 Lottery tickets (LTs) is able to discover accurate and sparse subnetworks that could be trained in isolation to match the performance of dense networks. Ensemble, in parallel, is one of the oldest time-proven tricks in machine learning to improve performance by combining the output of multiple independent models. However, the benefits of ensemble in the context of LTs will be diluted since ensemble does not directly lead to stronger sparse subnetworks, but leverages their predictions for a better decision. In this work, we first observe that directly averaging the weights of the adjacent learned subnetworks significantly boosts the performance of LTs. Encouraged by this observation, we further propose an alternative way to perform an ``ensemble'' over the subnetworks identified by iterative magnitude pruning via a simple interpolating strategy. We call our method "Lottery Pools". In contrast to the naive ensemble which brings no performance gains to each single subnetwork, Lottery Pools yields much stronger sparse subnetworks than the original LTs without requiring any extra training or inference cost. Across various modern architectures on CIFAR-10/100 and ImageNet, we show that our method achieves significant performance gains in both, in-distribution and out-of-distribution scenarios. Impressively, evaluated with VGG-16 and ResNet-18, the produced sparse subnetworks outperform the original LTs by up to 1.88% on CIFAR-100 and 2.36\% on CIFAR-100-C; the resulting dense network surpasses the pre-trained dense-model up to 2.22% on CIFAR-100 and 2.38% on CIFAR-100-C.

poster

In many scenarios of black-box optimization, evaluating the objective function values of solutions is expensive, while comparing a pair of solutions is relatively cheap, which yields the dueling black-box optimization. The side effect of dueling optimization is that it doubles the dimension of solution space and exacerbates the dimensionality scalability issue of black-box optimization, e.g., Bayesian optimization. To address this issue, the existing dueling optimization methods fix one solution when dueling throughout the optimization process, but it may reduce their efficacy. Fortunately, it has been observed that, in recommendation systems, the dueling results are mainly determined by the latent human preferences. In this paper, we abstract this phenomenon as the preferential intrinsic dimension and inject it into the dueling Bayesian optimization, resulting in the preferential embedding dueling Bayesian optimization (PE-DBO). PE-DBO decouples optimization and pairwise comparison via the preferential embedding matrix. Optimization is performed in the preferential intrinsic subspace with much lower dimensionality, while pairwise comparison is completed in the original dueling solution space. Theoretically, we disclose that the preference function can be approximately preserved in the lower-dimensional preferential intrinsic subspace. Experiment results verify that, on molecule discovery and web page recommendation dueling optimization tasks, the preferential intrinsic dimension exists and PE-DBO is superior in scalability compared with that of the state-of-the-art (SOTA) methods.

High-Dimensional Dueling Optimization with Preference Embedding

Attack Ensemble (AE), which combines multiple attacks together, provides a reliable way to evaluate adversarial robustness. In practice, AEs are often constructed and tuned by human experts, which however tends to be sub-optimal and time-consuming. In this work, we present AutoAE, a conceptually simple approach for automatically constructing AEs. In brief, AutoAE repeatedly adds the attack and its iteration steps to the ensemble that maximizes ensemble improvement per additional iteration consumed. We show theoretically that AutoAE yields AEs provably within a constant factor of the optimal for a given defense. We then use AutoAE to construct two AEs for $l_{\infty}$ and $l_2$ attacks, and apply them without any tuning or adaptation to 45 top adversarial defenses on the RobustBench leaderboard. In all except one cases we achieve equal or better (often the latter) robustness evaluation than existing AEs, and notably, in 29 cases we achieve better robustness evaluation than the best known one. Such performance of AutoAE shows itself as a reliable evaluation protocol for adversarial robustness, which further indicates the huge potential of automatic AE construction. Code is available at https://github.com/LeegerPENG/AutoAE.

Reliable Robustness Evaluation via Automatically Constructed Attack Ensembles

The spiking neural network (SNN) using leaky-integrated-and-fire (LIF) neurons has been commonly used in automatic speech recognition (ASR) tasks. However, the LIF neuron is still relatively simple compared to that in the biological brain. Further research on more types of neurons with different scales of neuronal dynamics is necessary. Here we introduce four types of neuronal dynamics to post-process the sequential patterns generated from the spiking transformer to get the complex dynamic neuron improved spiking transformer neural network (DyTr-SNN). We found that the DyTr-SNN could handle the non-toy automatic speech recognition task well, representing a lower phoneme error rate, lower computational cost, and higher robustness. These results indicate that the further cooperation of SNNs and neural dynamics at the neuron and network scales might have much in store for the future, especially on the ASR tasks.

Complex dynamic neurons improved spiking transformer network for efficient automatic speech recognition

The multi-view data with incomplete information hinder the effective data analysis. Existing multi-view imputation methods that learn the mapping between complete view and completely missing view are not able to deal with the common multi-view data with missing feature information. In this paper, we propose a generative imputation model named Git with optimal transport theory to jointly impute the missing features/values, conditional on all observed values from the multi-view data. Git consists of two modules, i.e., a multi-view joint generator (MJG) and a masking energy discriminator (MED). The generator MJG incorporates a joint autoencoder with the multiple imputation rule to learn the data distribution from all observed multi-view data. The discriminator MED leverages a new masking energy divergence function to make Git differentiable for imputation enhancement. Extensive experiments on several real-world multi-view data sets demonstrate that, Git yields over 35% accuracy gain, compared to the state-of-the-art approaches.

Jointly Imputing Multi-View Data with Optimal Transport

Optimal transport (OT) compares probability distributions by computing a meaningful alignment between their samples. CO-optimal transport (COOT) takes this comparison further by inferring an alignment between features as well. While this approach leads to better alignments and generalizes both OT and Gromov-Wasserstein distances, we provide a theoretical result showing that it is sensitive to outliers that are omnipresent in real-world data. This prompts us to propose unbalanced COOT for which we provably show its robustness to noise in the compared datasets. To the best of our knowledge, this is the first such result for OT methods in incomparable spaces. With this result in hand, we provide empirical evidence of this robustness for the challenging tasks of heterogeneous domain adaptation with and without varying proportions of classes and simultaneous alignment of samples and features across two single-cell measurements.

Unbalanced CO-Optimal Transport

We consider the problem of creating assistants that can help agents solve new sequential decision problems, assuming the agent is not able to specify the reward function explicitly to the assistant. Instead of acting in place of the agent as in current automation-based approaches, we give the assistant an advisory role and keep the agent in the loop as the main decision maker. The difficulty is that we must account for potential biases of the agent which may cause it to seemingly irrationally reject advice. To do this we introduce a novel formalization of assistance that models these biases, allowing the assistant to infer and adapt to them. We then introduce a new method for planning the assistant's actions which can scale to large decision making problems. We show experimentally that our approach adapts to these agent biases, and results in higher cumulative reward for the agent than automation-based alternatives. Lastly, we show that an approach combining advice and automation outperforms advice alone at the cost of losing some safety guarantees.

Zero-Shot Assistance in Sequential Decision Problems

Federated learning (FL) is a popular distributed machine learning paradigm which enables jointly training a global model without sharing clients' data. However, its repetitive server-client communication gives room for possible backdoor attacks which aims to mislead the global model into a targeted misprediction when a specific trigger pattern is presented. In response to such backdoor threats on federated learning, various defense measures have been proposed. In this paper, we study whether the current defense mechanisms truly neutralize the backdoor threats from federated learning in a practical setting by proposing a new federated backdoor attack framework for possible countermeasures. Different from traditional training (on triggered data) and rescaling (the malicious client model) based backdoor injection, the proposed backdoor attack framework (1) directly modifies (a small proportion of) local model weights to inject the backdoor trigger via sign flips; (2) jointly optimize the trigger pattern with the client model,  thus is more persistent and stealthy for circumventing existing defenses. In a case study, we examine the strength and weaknesses of several recent federated backdoor defenses from three major categories and provide suggestions to the practitioners when training federated models in practice.

On the Vulnerability of Backdoor Defenses for Federated Learning

Local graph neighborhood sampling is a fundamental computational problem that is at the heart of algorithms for node representation learning. 
Several works have presented algorithms for learning discrete node embeddings where graph nodes are represented by discrete features such as attributes of neighborhood nodes. Discrete embeddings offer several advantages compared to continuous word2vec-like node embeddings: ease of computation, scalability, and interpretability. We present LoNe Sampler, a suite of algorithms for generating discrete node embeddings by Local Neighborhood Sampling, and address two shortcomings of previous work. First, our algorithms have rigorously understood theoretical properties. Second, we show how to generate approximate explicit vector maps that avoid the expensive computation of a Gram matrix for the training of a kernel model. Experiments on benchmark datasets confirm the theoretical findings and demonstrate the advantages of the proposed methods. 

LoNe Sampler: Graph node embeddings by coordinated local neighborhood sampling

Recent investigations on rotation invariance for 3D point clouds have been devoted to devising rotation-invariant feature descriptors or learning canonical spaces where objects are semantically aligned.
Examinations of learning frameworks for invariance have seldom been looked into.
In this work, we review rotation invariance (RI) in terms of point cloud registration (PCR) and propose an effective framework for rotation invariance learning via three sequential stages, namely rotation-invariant shape encoding, aligned feature integration, and deep feature registration.
We first encode shape descriptors constructed with respect to reference frames defined over different scales, e.g., local patches and global topology, to generate rotation-invariant latent shape codes.
Within the integration stage, we propose an Aligned Integration Transformer (AIT) to produce a discriminative feature representation by integrating point-wise self- and cross-relations established within the shape codes.
Meanwhile, we adopt rigid transformations between reference frames to align the shape codes for feature consistency across different scales.
Finally, the deep integrated feature is registered to both rotation-invariant shape codes to maximize their feature similarities, such that rotation invariance of the integrated feature is preserved and shared semantic information is implicitly extracted from shape codes.
Experimental results on 3D shape classification, part segmentation, and retrieval tasks prove the feasibility of our framework.

Rethinking Rotation Invariance with Point Cloud Registration

When inferring reward functions from human behavior (be it demonstrations, comparisons, physical corrections, or e-stops), it has proven useful to model the human as making noisy-rational choices, with a ``rationality coefficient" capturing how much noise or entropy we expect to see in the human behavior. Prior work typically sets the rationality level to a constant value, regardless of the type, or quality, of human feedback. However, in many settings, giving one type of feedback (e.g. a demonstration) may be much more difficult than a different type of feedback (e.g. answering a comparison query). Thus, we expect to see more or less noise or suboptimality depending on the type of human feedback. In this work, we advocate that grounding the rationality coefficient in real data for each feedback type, rather than assuming a default value, has a significant positive effect on reward learning. We test this in both simulated experiments and in a user study with real human feedback. We find that overestimating human rationality can have dire effects on reward learning accuracy and regret. We also find that fitting the rationality coefficient to human data enables better reward learning, even when the human deviates significantly from the noisy-rational choice model due to systematic biases. Further, we find that the rationality level affects the informativeness of each feedback type: surprisingly, demonstrations are not always the most informative---when the human acts very suboptimally, comparisons actually become more informative, even when the rationality level is the same for both.  Ultimately, our results emphasize the importance and advantage of paying attention to the assumed human-rationality-level, especially when agents actively learn from multiple types of human feedback.

Premium content

Lottery Pools: Winning More by Interpolating Tickets without Increasing Training or Inference Cost

Downloads

Next from AAAI 2023

High-Dimensional Dueling Optimization with Preference Embedding

Similar lecture

A Tale of Two Latent Flows: Learning Latent Space Normalizing Flow with Short-run Langevin Flow for Approximate Inference