Search

Implicit discourse relation recognition involves determining relationships that hold between spans of text that are not linked by an explicit discourse connective. In recent years, the pre-train, prompt, and predict paradigm has emerged as a promising approach for tackling this task. However, previous work solely relied on manual verbalizers for implicit discourse relation recognition, which suffer from issues of ambiguity and even incorrectness. To overcome these limitations, we leverage the prototypes that capture certain class-level semantic features and the hierarchical label structure for different classes as the verbalizer. We show that our method improves on competitive baselines. Besides, our proposed approach can be extended to enable zero-shot cross-lingual learning, facilitating the recognition of discourse relations in languages with scarce resources. These advancement validate the practicality and versatility of our approach in addressing the issues of implicit discourse relation recognition across different languages.

Randomly pivoted Cholesky (RPCholesky) is an algorithm for constructing a low-rank approximation of a positive-semidefinite matrix using a small number of columns. This paper develops an accelerated version of RPCholesky that employs block matrix computations and rejection sampling to efficiently simulate the execution of the original algorithm. For the task of approximating a kernel matrix, the accelerated algorithm can run over $40\times$ faster. The paper contains implementation details, theoretical guarantees, experiments on benchmark data sets, and an application to computational chemistry., Comment: 26 pages, 4 figures, 5 pages of supplementary material

Privacy-preserving voice conversion aims to remove only the attributes of speech audio that convey identity information, keeping other speech characteristics intact. This paper presents a mechanism for privacy-preserving voice conversion that allows controlling the leakage of identity-bearing information using adversarial information hiding. This enables a deliberate trade-off between maintaining source-speech characteristics and modification of speaker identity. As such, the approach improves on voice-conversion techniques like CycleGAN and StarGAN, which were not designed for privacy, meaning that converted speech may leak personal information in unpredictable ways. Our approach is also more flexible than ASR-TTS voice conversion pipelines, which by design discard all prosodic information linked to textual content. Evaluations show that the proposed system successfully modifies perceived speaker identity whilst well maintaining source lexical content., Comment: Accepted for publication in proceedings of 4th symposium on security and privacy in speech communication

Over the last decade, experimental measurements of electrostatic screening lengths in concentrated electrolytes have exceeded theoretical predictions by orders of magnitude. This disagreement has led to a paradigm in which such screening lengths are referred to as 'anomalous underscreening', while others - predominantly those predicted by theory and molecular simulation - are referred to as 'normal underscreening'. Herein we use discrete Fourier analysis of the radial charge density obtained from molecular dynamics simulations to reveal the origin of anomalous underscreening in concentrated electrolytes. Normal underscreening above the Kirkwood point arises from low-frequency decay modes of the electrostatic potential, while anomalous underscreening arises from high-frequency decay modes that are observed only at high concentrations. The screening length associated with a particular decay mode is in turn determined by the degree of short-range interference between ion-ion correlation functions. The long-range decay associated with anomalous underscreening is thus ultimately determined by short range structure in the bulk electrolyte. These results reconcile the disagreement between experimental measurements and theoretical predictions of screening lengths in concentrated electrolytes., Comment: 16 pages, 3 figures

Modern platforms for potential qubit candidates, such as trapped ions or neutral atoms, allow long range connectivity between distant physical qubits through shuttling. This opens up an avenue for transversal logical CNOT gates between distant logical qubits, whereby physical CNOT gates are performed between each corresponding physical qubit on the control and target logical qubits. However, the transversal CNOT can propagate errors from one logical qubit to another, leading to correlated errors between logical qubits. We have developed a multi-pass iterative decoder that decodes each logical qubit separately to deal with this correlated error. We show that under circuit-level noise and only $\mathcal{O}(1)$ code cycles, a threshold can still persist, and the logical error rate will not be significantly degraded, matching the sub-threshold logical error rate scaling of $p^{\lfloor\frac{d}{2}\rfloor}$ for a distance $d$ rotated surface code.

Behavioural biometric authentication systems entail an enrolment period that is burdensome for the user. In this work, we explore generating synthetic gestures from a few real user gestures with generative deep learning, with the application of training a simple (i.e. non-deep-learned) authentication model. Specifically, we show that utilising synthetic data alongside real data can reduce the number of real datapoints a user must provide to enrol into a biometric system. To validate our methods, we use the publicly available dataset of WatchAuth, a system proposed in 2022 for authenticating smartwatch payments using the physical gesture of reaching towards a payment terminal. We develop a regularised autoencoder model for generating synthetic user-specific wrist motion data representing these physical gestures, and demonstrate the diversity and fidelity of our synthetic gestures. We show that using synthetic gestures in training can improve classification ability for a real-world system. Through this technique we can reduce the number of gestures required to enrol a user into a WatchAuth-like system by more than 40% without negatively impacting its error rates.

Markov state modeling has gained popularity in various scientific fields due to its ability to reduce complex time series data into transitions between a few states. Yet, current frameworks are limited by assuming a single Markov chain describes the data, and they suffer an inability to discern heterogeneities. As a solution, this paper proposes a variational expectation-maximization algorithm that identifies a mixture of Markov chains in a time-series data set. The method is agnostic to the definition of the Markov states, whether data-driven (e.g. by spectral clustering) or based on domain knowledge. Variational EM efficiently and organically identifies the number of Markov chains and dynamics of each chain without expensive model comparisons or posterior sampling. The approach is supported by a theoretical analysis and numerical experiments, including simulated and observational data sets based on ${\tt Last.fm}$ music listening, ultramarathon running, and gene expression. The results show the new algorithm is competitive with contemporary mixture modeling approaches and powerful in identifying meaningful heterogeneities in time series data.

Discourse relations play a pivotal role in establishing coherence within textual content, uniting sentences and clauses into a cohesive narrative. The Penn Discourse Treebank (PDTB) stands as one of the most extensively utilized datasets in this domain. In PDTB-3, the annotators can assign multiple labels to an example, when they believe that multiple relations are present. Prior research in discourse relation recognition has treated these instances as separate examples during training, and only one example needs to have its label predicted correctly for the instance to be judged as correct. However, this approach is inadequate, as it fails to account for the interdependence of labels in real-world contexts and to distinguish between cases where only one sense relation holds and cases where multiple relations hold simultaneously. In our work, we address this challenge by exploring various multi-label classification frameworks to handle implicit discourse relation recognition. We show that multi-label classification methods don't depress performance for single-label prediction. Additionally, we give comprehensive analysis of results and data. Our work contributes to advancing the understanding and application of discourse relations and provide a foundation for the future study, Comment: ACL2024 Finding

Superlatives are used to single out elements with a maximal/minimal property. Semantically, superlatives perform a set comparison: something (or some things) has the min/max property out of a set. As such, superlatives provide an ideal phenomenon for studying implicit phenomena and discourse restrictions. While this comparison set is often not explicitly defined, its (implicit) restrictions can be inferred from the discourse context the expression appears in. In this work we provide an extensive computational study on the semantics of superlatives. We propose a unified account of superlative semantics which allows us to derive a broad-coverage annotation schema. Using this unified schema we annotated a multi-domain dataset of superlatives and their semantic interpretations. We specifically focus on interpreting implicit or ambiguous superlative expressions, by analyzing how the discourse context restricts the set of interpretations. In a set of experiments we then analyze how well models perform at variations of predicting superlative semantics, with and without context. We show that the fine-grained semantics of superlatives in context can be challenging for contemporary models, including GPT-4., Comment: 11 pages

In the study of stochastic systems, the committor function describes the probability that a system starting from an initial configuration $x$ will reach a set $B$ before a set $A$. This paper introduces an efficient and interpretable algorithm for approximating the committor, called the "fast committor machine" (FCM). The FCM uses simulated trajectory data to build a kernel-based model of the committor. The kernel function is constructed to emphasize low-dimensional subspaces that optimally describe the $A$ to $B$ transitions. The coefficients in the kernel model are determined using randomized linear algebra, leading to a runtime that scales linearly in the number of data points. In numerical experiments involving a triple-well potential and alanine dipeptide, the FCM yields higher accuracy and trains more quickly than a neural network with the same number of parameters. The FCM is also more interpretable than the neural net., Comment: 10 pages, 7 figures

This paper develops a numerical procedure to accelerate the convergence of the Favre-averaged Non-Linear Harmonic (FNLH) method. The scheme provides a unified mathematical framework for solving the sparse linear systems formed by the mean flow and the time-linearized harmonic flows of FNLH in an explicit or implicit fashion. The approach explores the similarity of the sparse linear systems of FNLH and leads to a memory efficient procedure, so that its memory consumption does not depend on the number of harmonics to compute. The proposed method has been implemented in the industrial CFD solver HYDRA. Two test cases are used to conduct a comparative study of explicit and implicit schemes in terms of convergence, computational efficiency, and memory consumption. Comparisons show that the implicit scheme yields better convergence than the explicit scheme and is also roughly 7 to 10 times more computationally efficient than the explicit scheme with 4 levels of multigrid. Furthermore, the implicit scheme consumes only approximately $50\%$ of the explicit scheme with four levels of multigrid. Compared with the full annulus unsteady Reynolds averaged Navier-Stokes (URANS) simulations, the implicit scheme produces comparable results to URANS with computational time and memory consumption that are two orders of magnitude smaller.

In this paper, we investigate the reconstruction error, $N_\e^{\text{rec}}(x)$, when a linear, filtered back-projection (FBP) algorithm is applied to noisy, discrete Radon transform data with sampling step size $\epsilon$ in two-dimensions. Specifically, we analyze $N_\e^{\text{rec}}(x)$ for $x$ in small, $O(\e)$-sized neighborhoods around a generic fixed point, $x_0$, in the plane, where the measurement noise values, $\eta_{k,j}$ (i.e., the errors in the sinogram space), are random variables. The latter are independent, but not necessarily identically distributed. We show, under suitable assumptions on the first three moments of the $\eta_{k,j}$, that the following limit exists: $N^{\text{rec}}(\chx;x_0) = \lim_{\e\to0}N_\e^{\text{rec}}(x_0+\e\chx)$, for $\check x$ in a bounded domain. Here, $N_\e^{\text{rec}}$ and $ N^{\text{rec}}$ are viewed as continuous random variables, and the limit is understood in the sense of distributions. Once the limit is established, we prove that $N^{\text{rec}}$ is a zero mean Gaussian random field and compute explicitly its covariance. In addition, we validate our theory using numerical simulations and pseudo random noise.

Context. The third Gaia Data Release, which includes BP/RP spectra for 219 million sources, has opened a new window in the exploration of the chemical history and evolution of the Milky Way. The wealth of information encapsulated in these data is far greater than their low resolving power (R=50) at first glance would suggest, as shown in many studies. We zero in on the use of this data for the purpose of the detection of ''new'' metal-poor stars, which are hard to find yet essential for understanding - among other - several aspects of the origin of the Galaxy, star formation and the creation of the elements. Aims. We strive to refine a metal-poor candidate selection method which was developed with simulated Gaia BP/RP spectra, with an ultimate objective of providing the community with both a recipe to select stars for medium/high resolution observations and a catalogue of stellar metallicities. Methods. We used a datased comprised of GALAH DR3 and SAGA database stars in order to verify and adjust to real world data our selection method. For that purpose, we used dereddening as a mean to tackle the issue of extinction, and then we applied our fine-tuned method to select metal-poor candidates, which we thereafter observed and analysed. Results. We were able to infer metallicities for GALAH DR3 and SAGA stars - with color excesses up to E(B-V)<1.5 - with an uncertainty of 0.36 dex, which is good enough for the purpose of identifying new metal-poor stars. Further, we selected 26 metal-poor candidates - via our method - for observations. As spectral analysis showed, 100% of them had [Fe/H]<-2.0, 57% had [Fe/H]<-2.5 and 8% had [Fe/H]<-3.0. We inferred metallicities for these stars with an uncertainty of 0.31 dex, as was proven when comparing to the spectroscopic [Fe/H]. Finally, we assembled a catalogue of metallicities for 10 861 062 stars., Comment: 9 pages, 11 figures, accepted for publication in A&A

This analysis investigates energy performance of the residential and commercial building sectors in the Electric Reliability Council of Texas (ERCOT) during Winter Storm Uri. ERCOT electricity demand was modeled for the ERCOT baseline building stock as well as for the baseline building stock retrofitted with an efficiency upgrade package, an electrification upgrade package, and an efficiency + electrification upgrade package. The electrification scenario that retrofitted buildings with air-source heat pumps (ASHPs) would have lowered ERCOT daily peak electricity demand relative to the baseline scenario for every day of the year, except during the week of Winter Storm Uri. As the mean outdoor temperature dropped below -5{\deg}C (23{\deg}F), diminishing ASHP efficiency would have resulted in electrification scenario demand exceeding the two distinct baseline scenario daily demand peaks on February 15th and 16th (87.3 GW and 88.7 GW) to hit 111.8 GW and 117.5 GW. The efficiency package would have lowered daily peak demand on these days to 67.0 GW and 68.0 GW. The efficiency + electrification package would have lowered peak demand on these days to 81.5 GW and 85.6 GW. When electricity shortfall profiles were produced by comparing modeled electricity demand to actual ERCOT electricity generation during the storm, the results indicate that the electrification scenario electricity shortfall (1741 GWh) would have been larger than for the baseline scenario (1225 GWh) and the electricity shortfalls for the efficiency scenario (347 GWh) and efficiency + electrification scenario (704 GWh) would have been lower than the baseline. The efficiency, electrification, and efficiency + electrification scenarios would all have lowered summer daily peak demand due to improvements in building cooling efficiency and would have lowered annual electricity consumption by 5.9%, 6.8%, and 11.9%, respectively., Comment: 35 pages, 13 figures

Rare event schemes require an approximation of the probability of the rare event as a function of system state. Finding an appropriate reaction coordinate is typically the most challenging aspect of applying a rare event scheme. Here we develop an artificial intelligence (AI) based reaction coordinate that effectively predicts which of a limited number of simulations of the Solar System will go unstable using a convolutional neural network classifier. The performance of the algorithm does not degrade significantly even 3.5 billion years before the instability. We overcome the class imbalance intrinsic to rare event problems using a combination of minority class oversampling, increased minority class weighting, and pulling multiple non-overlapping training sequences from simulations. Our success suggests that AI may provide a promising avenue for developing reaction coordinates without detailed theoretical knowledge of the system., Comment: 3 pages, 1 figure