Distribution alignment can be used to learn invariant representations with applications in fairness and robustness. Most prior works resort to adversarial alignment methods but the resulting minimax problems are unstable and challenging to optimize. Non-adversarial likelihood-based approaches either require model invertibility, impose constraints on the latent prior, or lack a generic framework for alignment. To overcome these limitations, we propose a non-adversarial VAE-based alignment method that can be applied to any model pipeline. We develop a set of alignment upper bounds (including a noisy bound) that have VAE-like objectives but with a different perspective. We carefully compare our method to prior VAE-based alignment approaches both theoretically and empirically. Finally, we demonstrate that our novel alignment losses can replace adversarial losses in standard invariant representation learning pipelines without modifying the original architectures -- thereby significantly broadening the applicability of non-adversarial alignment methods.
相關內容
Numerical solution of discrete PDEs corresponding to saddle point problems is highly relevant to physical systems such as Stokes flow. However, scaling up numerical solvers for such systems is often met with challenges in efficiency and convergence. Multigrid is an approach with excellent applicability to elliptic problems such as the Stokes equations, and can be a solution to such challenges of scalability and efficiency. The degree of success of such methods, however, is highly contingent on the design of key components of a multigrid scheme, including the hierarchy of discretizations, and the relaxation scheme used. Additionally, in many practical cases, it may be more effective to use a multigrid scheme as a preconditioner to an iterative Krylov subspace solver, as opposed to striving for maximum efficacy of the relaxation scheme in all foreseeable settings. In this paper, we propose an efficient symmetric multigrid preconditioner for the Stokes Equations on a staggered finite-difference discretization. Our contribution is focused on crafting a preconditioner that (a) is symmetric indefinite, matching the property of the Stokes system itself, (b) is appropriate for preconditioning the SQMR iterative scheme, and (c) has the requisite symmetry properties to be used in this context. In addition, our design is efficient in terms of computational cost and facilitates scaling to large domains.
Receive generalized spatial modulation (RGSM), as an advanced type of receive spatial modulation (RSM), can be divided into diversity and multiplexing (MUX) schemes according to whether the symbols received by the selected antennas are the same. Recently, reconfigurable intelligent surface (RIS) assisted RSM has attracted much attention due to better reception performance and spectral efficiency. The RIS-aided RGSM (RIS-RGSM) with diversity scheme is realized in this paper via a simple improvement based on the state-of-the-art RIS-aided receive generalized space shift keying (RIS-RGSSK) scheme. To increase the transmission rate, the RIS-RGSM with MUX scheme is proposed in this paper, which adjusts the reflecting phase shifts and on/off states of RIS elements. Theoretical analysis and numerical simulations show that the proposed RIS-RGSM with MUX scheme has better bit error rate (BER) performance than the diversity scheme. Compared to the RIS-RGSSK scheme, the proposed RIS-RGSM scheme can significantly reduce the number of receive antennas while maintaining the transmission rate.
While coresets have been growing in terms of their application, barring few exceptions, they have mostly been limited to unsupervised settings. We consider supervised classification problems, and non-decomposable evaluation measures in such settings. We show that stratified uniform sampling based coresets have excellent empirical performance that are backed by theoretical guarantees too. We focus on the F1 score and Matthews Correlation Coefficient, two widely used non-decomposable objective functions that are nontrivial to optimize for and show that uniform coresets attain a lower bound for coreset size, and have good empirical performance, comparable with ``smarter'' coreset construction strategies.
Efficiently computing spatio-textual queries has become increasingly important in various applications that need to quickly retrieve geolocated entities associated with textual information, such as in location-based services and social networks. To accelerate such queries, several works have proposed combining spatial and textual indices into hybrid index structures. Recently, the novel idea of replacing traditional indices with ML models has attracted a lot of attention. This includes works on learned spatial indices, where the main challenge is to address the lack of a total ordering among objects in a multidimensional space. In this work, we investigate how to extend this novel type of index design to the case of spatio-textual data. We study different design choices, based on either loose or tight coupling between the spatial and textual part, as well as a hybrid index that combines a traditional and a learned component. We also perform an experimental evaluation using several real-world datasets to assess the potential benefits of using a learned index for evaluating spatio-textual queries.
The primary contribution of this paper is new methods for reducing communication in the sampling step for distributed GNN training. Here, we propose a matrix-based bulk sampling approach that expresses sampling as a sparse matrix multiplication (SpGEMM) and samples multiple minibatches at once. When the input graph topology does not fit on a single device, our method distributes the graph and use communication-avoiding SpGEMM algorithms to scale GNN minibatch sampling, enabling GNN training on much larger graphs than those that can fit into a single device memory. When the input graph topology (but not the embeddings) fits in the memory of one GPU, our approach (1) performs sampling without communication, (2) amortizes the overheads of sampling a minibatch, and (3) can represent multiple sampling algorithms by simply using different matrix constructions. In addition to new methods for sampling, we show that judiciously replicating feature data with a simple all-to-all exchange can outperform current methods for the feature extraction step in distributed GNN training. We provide experimental results on the largest Open Graph Benchmark (OGB) datasets on $128$ GPUs, and show that our pipeline is $2.5\times$ faster Quiver (a distributed extension to PyTorch-Geometric) on a $3$-layer GraphSAGE network. On datasets outside of OGB, we show a $8.46\times$ speedup on $128$ GPUs in-per epoch time. Finally, we show scaling when the graph is distributed across GPUs and scaling for both node-wise and layer-wise sampling algorithms
This paper presents a motion planning algorithm for quadruped locomotion based on density functions. We decompose the locomotion problem into a high-level density planner and a model predictive controller (MPC). Due to density functions having a physical interpretation through the notion of occupancy, it is intuitive to represent the environment with safety constraints. Hence, there is an ease of use to constructing the planning problem with density. The proposed method uses a simplified model of the robot into an integrator system, where the high-level plan is in a feedback form formulated through an analytically constructed density function. We then use the MPC to optimize the reference trajectory, in which a low-level PID controller is used to obtain the torque level control. The overall framework is implemented in simulation, demonstrating our feedback density planner for legged locomotion. The implementation of work is available at \url{//github.com/AndrewZheng-1011/legged_planner}
The prevalence of the powerful multilingual models, such as Whisper, has significantly advanced the researches on speech recognition. However, these models often struggle with handling the code-switching setting, which is essential in multilingual speech recognition. Recent studies have attempted to address this setting by separating the modules for different languages to ensure distinct latent representations for languages. Some other methods considered the switching mechanism based on language identification. In this study, a new attention-guided adaptation is proposed to conduct parameter-efficient learning for bilingual ASR. This method selects those attention heads in a model which closely express language identities and then guided those heads to be correctly attended with their corresponding languages. The experiments on the Mandarin-English code-switching speech corpus show that the proposed approach achieves a 14.2% mixed error rate, surpassing state-of-the-art method, where only 5.6% additional parameters over Whisper are trained.
Dialogue systems are frequently updated to accommodate new services, but naively updating them by continually training with data for new services in diminishing performance on previously learnt services. Motivated by the insight that dialogue state tracking (DST), a crucial component of dialogue systems that estimates the user's goal as a conversation proceeds, is a simple natural language understanding task, we propose reformulating it as a bundle of granular example-guided question answering tasks to minimize the task shift between services and thus benefit continual learning. Our approach alleviates service-specific memorization and teaches a model to contextualize the given question and example to extract the necessary information from the conversation. We find that a model with just 60M parameters can achieve a significant boost by learning to learn from in-context examples retrieved by a retriever trained to identify turns with similar dialogue state changes. Combining our method with dialogue-level memory replay, our approach attains state of the art performance on DST continual learning metrics without relying on any complex regularization or parameter expansion methods.
This manuscript portrays optimization as a process. In many practical applications the environment is so complex that it is infeasible to lay out a comprehensive theoretical model and use classical algorithmic theory and mathematical optimization. It is necessary as well as beneficial to take a robust approach, by applying an optimization method that learns as one goes along, learning from experience as more aspects of the problem are observed. This view of optimization as a process has become prominent in varied fields and has led to some spectacular success in modeling and systems that are now part of our daily lives.
Knowledge graphs (KGs) serve as useful resources for various natural language processing applications. Previous KG completion approaches require a large number of training instances (i.e., head-tail entity pairs) for every relation. The real case is that for most of the relations, very few entity pairs are available. Existing work of one-shot learning limits method generalizability for few-shot scenarios and does not fully use the supervisory information; however, few-shot KG completion has not been well studied yet. In this work, we propose a novel few-shot relation learning model (FSRL) that aims at discovering facts of new relations with few-shot references. FSRL can effectively capture knowledge from heterogeneous graph structure, aggregate representations of few-shot references, and match similar entity pairs of reference set for every relation. Extensive experiments on two public datasets demonstrate that FSRL outperforms the state-of-the-art.
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191