Speaker individuality information is among the most critical elements within speech signals. By thoroughly and accurately modeling this information, it can be utilized in various intelligent speech applications, such as speaker recognition, speaker diarization, speech synthesis, and target speaker extraction. In this overview, we present a comprehensive review of neural approaches to speaker representation learning from both theoretical and practical perspectives. Theoretically, we discuss speaker encoders ranging from supervised to self-supervised learning algorithms, standalone models to large pretrained models, pure speaker embedding learning to joint optimization with downstream tasks, and efforts toward interpretability. Practically, we systematically examine approaches for robustness and effectiveness, introduce and compare various open-source toolkits in the field. Through the systematic and comprehensive review of the relevant literature, research activities, and resources, we provide a clear reference for researchers in the speaker characterization and modeling field, as well as for those who wish to apply speaker modeling techniques to specific downstream tasks.
Neural scaling laws are observed in a range of domains, to date with no clear understanding of why they occur. Recent theories suggest that loss power laws arise from Zipf's law, a power law observed in domains like natural language. One theory suggests that language scaling laws emerge when Zipf-distributed task quanta are learned in descending order of frequency. In this paper we examine power-law scaling in AlphaZero, a reinforcement learning algorithm, using a theory of language-model scaling. We find that game states in training and inference data scale with Zipf's law, which is known to arise from the tree structure of the environment, and examine the correlation between scaling-law and Zipf's-law exponents. In agreement with quanta scaling theory, we find that agents optimize state loss in descending order of frequency, even though this order scales inversely with modelling complexity. We also find that inverse scaling, the failure of models to improve with size, is correlated with unusual Zipf curves where end-game states are among the most frequent states. We show evidence that larger models shift their focus to these less-important states, sacrificing their understanding of important early-game states.
Use real word data to evaluate the performance of the electrocardiographic markers of GEH as features in a machine learning model with Standard ECG features and Risk Factors in Predicting Outcome of patients in a population referred to a tertiary cardiology hospital. Patients forwarded to specific evaluation in a cardiology specialized hospital performed an ECG and a risk factor anamnesis. A series of follow up attendances occurred in periods of 6 months, 12 months and 15 months to check for cardiovascular related events (mortality or new nonfatal cardiovascular events (Stroke, MI, PCI, CS), as identified during 1-year phone follow-ups. The first attendance ECG was measured by a specialist and processed in order to obtain the global electric heterogeneity (GEH) using the Kors Matriz. The ECG measurements, GEH parameters and risk factors were combined for training multiple instances of XGBoost decision trees models. Each instance were optmized for the AUCPR and the instance with higher AUC is chosen as representative to the model. The importance of each parameter for the winner tree model was compared to better understand the improvement from using GEH parameters. The GEH parameters turned out to have statistical significance for this population specially the QRST angle and the SVG. The combined model with the tree parameters class had the best performance. The findings suggest that using VCG features can facilitate more accurate identification of patients who require tertiary care, thereby optimizing resource allocation and improving patient outcomes. Moreover, the decision tree model's transparency and ability to pinpoint critical features make it a valuable tool for clinical decision-making and align well with existing clinical practices.
Population protocols are a model of distributed computation in which an arbitrary number of indistinguishable finite-state agents interact in pairs to decide some property of their initial configuration. We investigate the behaviour of population protocols under adversarial faults that cause agents to silently crash and no longer interact with other agents. As a starting point, we consider the property ``the number of agents exceeds a given threshold $t$'', represented by the predicate $x \geq t$, and show that the standard protocol for $x \geq t$ is very fragile: one single crash in a computation with $x:=2t-1$ agents can already cause the protocol to answer incorrectly that $x \geq t$ does not hold. However, a slightly less known protocol is robust: for any number $t' \geq t$ of agents, at least $t' - t+1$ crashes must occur for the protocol to answer that the property does not hold. We formally define robustness for arbitrary population protocols, and investigate the question whether every predicate computable by population protocols has a robust protocol. Angluin et al. proved in 2007 that population protocols decide exactly the Presburger predicates, which can be represented as Boolean combinations of threshold predicates of the form $\sum_{i=1}^n a_i \cdot x_i \geq t$ for $a_1,...,a_n, t \in \mathbb{Z}$ and modulo prdicates of the form $\sum_{i=1}^n a_i \cdot x_i \bmod m \geq t $ for $a_1, \ldots, a_n, m, t \in \mathbb{N}$. We design robust protocols for all threshold and modulo predicates. We also show that, unfortunately, the techniques in the literature that construct a protocol for a Boolean combination of predicates given protocols for the conjuncts do not preserve robustness. So the question remains open.
Source detection (SD) is the task of finding the origin of a spreading process in a network. Algorithms for SD help us combat diseases, misinformation, pollution, and more, and have been studied by physicians, physicists, sociologists, and computer scientists. The field has received considerable attention and been analyzed in many settings (e.g., under different models of spreading processes), yet all previous work shares the same assumption that the network the spreading process takes place in has the same structure at every point in time. For example, if we consider how a disease spreads through a population, it is unrealistic to assume that two people can either never or at every time infect each other, rather such an infection is possible precisely when they meet. Therefore, we propose an extended model of SD based on temporal graphs, where each link between two nodes is only present at some time step. Temporal graphs have become a standard model of time-varying graphs, and, recently, researchers have begun to study infection problems (such as influence maximization) on temporal graphs (arXiv:2303.11703, [Gayraud et al., 2015]). We give the first formalization of SD on temporal graphs. For this, we employ the standard SIR model of spreading processes ([Hethcote, 1989]). We give both lower bounds and algorithms for the SD problem in a number of different settings, such as with consistent or dynamic source behavior and on general graphs as well as on trees.
Hardware accelerators like quantum annealers or neuromorphic chips are capable of finding the ground state of a Hamiltonian. A promising route in utilizing these devices is via methods from automated reasoning: The problem at hand is first encoded into MaxSAT; then MaxSAT is reduced to Max2SAT; and finally, Max2SAT is translated into a Hamiltonian. It was observed that different encodings can dramatically affect the efficiency of the hardware accelerators. Yet, previous studies were only concerned with the size of the encodings rather than with syntactic or structural properties. We establish structure-aware reductions between MaxSAT, Max2SAT, and the quadratic unconstrained binary optimization problem (QUBO) that underlies such hardware accelerators. All these problems turn out to be equivalent under linear-time, treewidth-preserving reductions. As a consequence, we obtain tight lower bounds under ETH and SETH for Max2SAT and QUBO, as well as a new time-optimal fixed-parameter algorithm for QUBO. While our results are tight up to a constant additive factor for the primal treewidth, we require a constant multiplicative factor for the incidence treewidth. To close the emerging gap, we supplement our results with novel time-optimal algorithms for fragments of MaxSAT based on model counting.
In the burgeoning field of medical imaging, precise computation of 3D volume holds a significant importance for subsequent qualitative analysis of 3D reconstructed objects. Combining multivariate calculus, marching cube algorithm, and binary indexed tree data structure, we developed an algorithm for efficient computation of intrinsic volume of any volumetric data recovered from computed tomography (CT) or magnetic resonance (MR). We proposed the 30 configurations of volume values based on the polygonal mesh generation method. Our algorithm processes the data in scan-line order simultaneously with reconstruction algorithm to create a Fenwick tree, ensuring query time much faster and assisting users' edition of slicing or transforming model. We tested the algorithm's accuracy on simple 3D objects (e.g., sphere, cylinder) to complicated structures (e.g., lungs, cardiac chambers). The result deviated within $\pm 0.004 \text{cm}^3$ and there is still room for further improvement.
Causal inference has shown potential in enhancing the predictive accuracy, fairness, robustness, and explainability of Natural Language Processing (NLP) models by capturing causal relationships among variables. The emergence of generative Large Language Models (LLMs) has significantly impacted various NLP domains, particularly through their advanced reasoning capabilities. This survey focuses on evaluating and improving LLMs from a causal view in the following areas: understanding and improving the LLMs' reasoning capacity, addressing fairness and safety issues in LLMs, complementing LLMs with explanations, and handling multimodality. Meanwhile, LLMs' strong reasoning capacities can in turn contribute to the field of causal inference by aiding causal relationship discovery and causal effect estimations. This review explores the interplay between causal inference frameworks and LLMs from both perspectives, emphasizing their collective potential to further the development of more advanced and equitable artificial intelligence systems.
Big models have achieved revolutionary breakthroughs in the field of AI, but they might also pose potential concerns. Addressing such concerns, alignment technologies were introduced to make these models conform to human preferences and values. Despite considerable advancements in the past year, various challenges lie in establishing the optimal alignment strategy, such as data cost and scalable oversight, and how to align remains an open question. In this survey paper, we comprehensively investigate value alignment approaches. We first unpack the historical context of alignment tracing back to the 1920s (where it comes from), then delve into the mathematical essence of alignment (what it is), shedding light on the inherent challenges. Following this foundation, we provide a detailed examination of existing alignment methods, which fall into three categories: Reinforcement Learning, Supervised Fine-Tuning, and In-context Learning, and demonstrate their intrinsic connections, strengths, and limitations, helping readers better understand this research area. In addition, two emerging topics, personal alignment, and multimodal alignment, are also discussed as novel frontiers in this field. Looking forward, we discuss potential alignment paradigms and how they could handle remaining challenges, prospecting where future alignment will go.
The advent of large language models marks a revolutionary breakthrough in artificial intelligence. With the unprecedented scale of training and model parameters, the capability of large language models has been dramatically improved, leading to human-like performances in understanding, language synthesizing, and common-sense reasoning, etc. Such a major leap-forward in general AI capacity will change the pattern of how personalization is conducted. For one thing, it will reform the way of interaction between humans and personalization systems. Instead of being a passive medium of information filtering, large language models present the foundation for active user engagement. On top of such a new foundation, user requests can be proactively explored, and user's required information can be delivered in a natural and explainable way. For another thing, it will also considerably expand the scope of personalization, making it grow from the sole function of collecting personalized information to the compound function of providing personalized services. By leveraging large language models as general-purpose interface, the personalization systems may compile user requests into plans, calls the functions of external tools to execute the plans, and integrate the tools' outputs to complete the end-to-end personalization tasks. Today, large language models are still being developed, whereas the application in personalization is largely unexplored. Therefore, we consider it to be the right time to review the challenges in personalization and the opportunities to address them with LLMs. In particular, we dedicate this perspective paper to the discussion of the following aspects: the development and challenges for the existing personalization system, the newly emerged capabilities of large language models, and the potential ways of making use of large language models for personalization.
While it is nearly effortless for humans to quickly assess the perceptual similarity between two images, the underlying processes are thought to be quite complex. Despite this, the most widely used perceptual metrics today, such as PSNR and SSIM, are simple, shallow functions, and fail to account for many nuances of human perception. Recently, the deep learning community has found that features of the VGG network trained on the ImageNet classification task has been remarkably useful as a training loss for image synthesis. But how perceptual are these so-called "perceptual losses"? What elements are critical for their success? To answer these questions, we introduce a new Full Reference Image Quality Assessment (FR-IQA) dataset of perceptual human judgments, orders of magnitude larger than previous datasets. We systematically evaluate deep features across different architectures and tasks and compare them with classic metrics. We find that deep features outperform all previous metrics by huge margins. More surprisingly, this result is not restricted to ImageNet-trained VGG features, but holds across different deep architectures and levels of supervision (supervised, self-supervised, or even unsupervised). Our results suggest that perceptual similarity is an emergent property shared across deep visual representations.