StrTransformer recovers sources without labels by optimizing latent matrices and enforcing structural constraints.
problem Unsupervised blind source recovery in signal processing.
method Source-wise structured Transformer framework with latent source matrix optimization, structural regularization, and branch-specific weights.
result StrTransformer learns distinct temporal-scale structures and recovers source-aligned latent trajectories.
New WildMix dataset and Spectro-Temporal Transformer model for better monoaural audio source separation.
problem Challenging monoaural audio source separation.
method Introducing WildMix dataset and Spectro-Temporal Transformer model (STT) with Spectro-Temporal Encoder (STE).
result STT outperforms previous baselines on the WildMix dataset.
Transformer model improves source code summarization.
problem Generating readable summaries of source code.
method Transformer model with self-attention mechanism for code representation.
result Transformer model outperforms state-of-the-art techniques.
Paper uses time series transformers to predict investment success.
problem Optimizing investment sourcing in VC and GC.
method Transformer-based Multivariate Time Series Classifier (TMTSC).
result TMTSC improves decision making in VC and GC investments.
Tree-Transformer improves grammar correction in code and natural language.
problem Grammar correction in tree-structured data.
method Tree-Transformer neural network architecture for tree-structured data.
result Significant improvement in grammar correction accuracy.
Transformers can learn Markov processes with constant depth, surprising results.
problem Understanding how transformers learn context in Markov processes.
method Empirical study and theoretical analysis of attention-based transformers on Markov data.
result Transformers with constant depth can achieve low test loss on Markov sequences, matching empirical and theoretical findings.
Proposes a framework to fuse heterogeneous data sources for better modeling.
problem Heterogeneous data sources with different input parameter spaces.
method Input mapping calibration (IMC) and latent variable Gaussian process (LVGP).
result Improved predictive accuracy over single source models.
CWAN tackles multi-source heterogeneous domain adaptation with conditional weighting.
problem Learning cross-domain samples from multiple heterogeneous domains.
method CWAN uses a feature transformer, label classifier, and domain discriminator to learn from multiple sources.
result CWAN outperforms state-of-the-art methods on four real-world datasets.
Paper presents a code authorship attribution attack using adversarial learning.
problem Misleading attribution of source code using machine learning methods.
method Exploits adversarial examples and semantics-preserving code transformations guided by Monte-Carlo tree search.
result Demonstrates substantial effect on attribution methods, reducing accuracy from over 88% to 1%.
SAHMM-VAE separates sources adaptively using hidden Markov priors.
problem Unsupervised blind source separation.
method Source-wise adaptive Hidden Markov prior variational autoencoder.
result Different latent dimensions align with different source-specific temporal organizations.
Transformer models show robustness across domains with domain adversarial training.
problem Domain adaptation from multiple sources with no labeled data.
method Domain adversarial training and mixture of experts.
result Domain adversarial training improves representation but not performance.
Survey explores methods to adapt deep learning models across multiple labeled domains.
problem Difficulty in obtaining labeled data for deep learning models.
method Multi-source domain adaptation (MDA) to transfer knowledge from labeled to unlabeled or sparsely labeled target domains.
result MDA methods improve performance by minimizing domain shift.
Transformers with MLP heads outperform linear baselines in in-context learning.
problem Understanding in-context learning in Transformers with nonlinear MLP heads.
method Analyzing a model with two-layer MLPs trained via gradient steps and fully optimized, under high-dimensional asymptotics.
result Nonlinear MLPs enhance ICL performance, particularly on nonlinear tasks.
Method converts facial expressions and voice of a source speaker into a target speaker.
problem Separate conversion of facial and acoustic features leads to unnatural results.
method Uses three neural networks: conversion, waveform generation, and image reconstruction.
result Significantly higher naturalness achieved when converting both features together.
NSF models generate speech waveforms faster and better than WaveNet.
problem Efficiently generating speech waveforms for statistical parametric synthesis.
method Neural source-filter (NSF) models that combine sine-based excitation, non-AR filter, and conditional preprocessing.
result NSF models generate waveforms 100 times faster than WaveNet and have better quality.
Study on flow-based methods for capturing tail properties in densities.
problem Flow-based methods struggle with capturing non-Gaussian tails.
method Characterize and adapt triangular maps to capture tail properties.
result Flow models lack the ability to capture non-Gaussian tails.
Efficiently transforms samples from various statistical models.
problem Approximately transforming samples from one statistical model to another without knowing the source model's parameters.
method Constructs computationally efficient procedures to reduce uniform, Erlang, and Laplace models to general target families.
result Establishes nonasymptotic reductions between canonical high-dimensional problems, such as mixtures of experts, phase retrieval, and signal denoising.
A new method for blind source separation using hierarchical structure and KL divergence.
problem Blind source separation of complex interacting signals.
method Hierarchical log-linear model with KL divergence minimization.
result Superior performance compared to existing techniques on images and time series data.
Non-negative blind source separation (non-negative BSS), which is also referred to as non-negative matrix factorization (NMF), is a very active field in domains as different as astrophysics, audio processing or biomedical signal processing. In this context, the efficient retrieval of the sources requires the use of sig…
New method reduces label and data shifts between domains using optimal transport.
problem Label shift between source and target domains in domain adaptation.
method Developed theory and LDROT method to mitigate label and data shifts.
result Theoretical and experimental validation of LDROT's effectiveness.
The need to efficiently calculate first- and higher-order derivatives of increasingly complex models expressed in Python has stressed or exceeded the capabilities of available tools. In this work, we explore techniques from the field of automatic differentiation (AD) that can give researchers expressive power, performa…
Proposes a two-step method for sound source separation.
problem Improving sound source separation performance.
method First, learn a latent space transform. Second, train a separation module in the latent space.
result The proposed method achieves better performance than joint learning approaches.
StonkBERT predicts stock price movements using company text data.
problem Can language models predict medium-run stock price movements?
method Fine-tuning transformer-based language models (BERT) on company text data (news articles, blogs, annual reports) for stock price performance classification.
result StonkBERT shows substantial improvement in predictive accuracy compared to traditional models, with news articles providing the best results.
MAGIC generates image collages from set templates using attention and set representations.
problem Generating image collages from set templates is challenging for classical models.
method Memory Attentive Generation of Image Collages (MAGIC) using Set-Transformer layers and set-pooling.
result MAGIC can generate image collages from set templates in one forward pass.
This paper proposes network recasting as a general method for network architecture transformation. The primary goal of this method is to accelerate the inference process through the transformation, but there can be many other practical applications. The method is based on block-wise recasting; it recasts each source bl…
Study makes code models robust to small changes that keep functionality.
problem Vulnerability of deep neural networks to adversarial examples in source code.
method Defined a powerful adversary and adversarial training to learn robust models.
result Significant gains in robustness demonstrated across different languages and architectures.
Automatic differentiation (AD) is an essential primitive for machine learning programming systems. Tangent is a new library that performs AD using source code transformation (SCT) in Python. It takes numeric functions written in a syntactic subset of Python and NumPy as input, and generates new Python functions which c…
Kymatio simplifies scattering transforms for Python.
problem Signal processing and machine learning applications.
method Wavelet scattering transform implemented in Python.
result Efficient, GPU-accelerated implementation.
The paper explores optimal algorithms for linear regression under covariate shift, proving the optimality of certain transformations and SGD variants.
problem Optimal algorithms for linear regression under covariate shift with ellipse-shaped constraints.
method Establishes a tight lower generalization bound via Bayesian Cramer-Rao inequality, proves the optimality of certain transformations, and analyzes SGD variants.
result Optimal estimators and SGD variants achieve optimality under specific conditions.
This study predicts parking availability using multi-source data and a self-supervised learning enhanced transformer.
problem Accurate parking availability prediction to support urban planning and management.
method Proposes SST-iTransformer, a self-supervised learning enhanced spatio-temporal inverted transformer, integrating multi-source data.
result SST-iTransformer achieves state-of-the-art performance in parking availability prediction.
Geometrically transforms word embeddings into a common space for better comparison.
problem Comparing embeddings from different sources is challenging.
method Applies orthogonal rotations and Mahalanobis scaling to transform embeddings into a shared latent space.
result The method improves word similarity and analogy tasks.
SIG model identifies invariant variables for MSDA with fewer domain constraints.
problem Challenges in enforcing minimal changes across domains for MSDA.
method Subspace identification theory and variational inference.
result SIG model outperforms existing techniques on various benchmark datasets.
We present a system for the automatic differentiation of a higher-order functional array-processing language. The core functional language underlying this system simultaneously supports both source-to-source automatic differentiation and global optimizations such as loop transformations. Thanks to this feature, we demo…
A new method matches measures across different spaces using cost-regularized optimal transport.
problem Matching measures in different spaces without aligned data.
method Cost-regularized optimal transport formulation to match measures across two Euclidean spaces.
result Demonstrated applicability to single-cell spatial transcriptomics/multiomics matching tasks.
Tracr compiles programs into transformer models for interpretability.
problem Uncertainty in understanding transformer model outputs due to unknown learned programs.
method Tracr compiles human-readable programs into known structure transformer models.
result Known structure of Tracr-compiled models serves as ground-truth for interpretability.
In this article, we consider a generalized Radon transform that comes up in ultrasound reflection tomography. In our model, the ultrasound emitter and receiver move at a constant distance apart along a circle. We analyze the microlocal properties of the transform R that arises from this model. As a consequence, we sh…
COSET benchmarks neural program embeddings using diverse source-code datasets.
problem Evaluating neural program embeddings is challenging due to lack of straightforward metrics.
method COSET framework with labeled programs, transformations, and a pilot study.
result COSET identifies strengths and weaknesses of neural models and program characteristics.
A novel framework synthesizes treatment data across sites using optimal transport.
problem Estimating treatment effects across different sites with varying conditions.
method Distributional causal inference, Optimal Transport for alignment of control group distributions.
result Synthetic treatment group data aligns with true target distribution under general conditions.
The paper proposes a method to identify power system oscillation modes using blind source separation.
problem Accurately identifying oscillation modes in power systems with renewable energy sources.
method A high-order blind source identification (HOBI) algorithm based on copula statistic combined with Hilbert transform and iteration procedure.
result The method can identify all oscillation modes and model order from a single channel of observation signals, outperforming state-of-the-art methods.
Tool visualizes Transformer model attention for better understanding.
problem Understanding complex attention mechanisms in deep learning models.
method Developed an open-source tool to visualize attention at three levels.
result Visualization helps interpret and analyze Transformer models.
We construct a form of swallowtail singularity in R^3 which uses coordinate transformation on the source and isometry on the target. As an application, we classify configurations of asymptotic curves and characteristic curves near swallowtail.
Myia compiler optimizes ML models with efficient AD for array programming.
problem Efficient automatic differentiation for array programming in ML.
method Introduces a new graph-based IR that supports function calls, higher-order functions, and recursion.
result Myia compiler enables efficient AD using source transformation without a tape, supporting higher-order derivatives.
We consider the Hypothesis Transfer Learning (HTL) problem where one incorporates a hypothesis trained on the source domain into the learning procedure of the target domain. Existing theoretical analysis either only studies specific algorithms or only presents upper bounds on the generalization error but not on the exc…
Efficient trainable front-end for neural speech enhancement.
problem Inefficient STFT front-ends in neural speech enhancement models.
method Butterfly mechanism for Fast Fourier Transform, trainable STFT window.
result Accuracy and efficiency improvements for low-compute systems.
A multi-step framework tackles online unsupervised domain adaptation with novel mean-target subspace computation.
problem Online unsupervised domain adaptation with unlabelled target data arriving sequentially.
method Multi-step framework with a novel mean-target subspace computation and temporal coherency consideration.
result Improved performance over previous approaches on four datasets.
Self-taught learning is a technique that uses a large number of unlabeled data as source samples to improve the task performance on target samples. Compared with other transfer learning techniques, self-taught learning can be applied to a broader set of scenarios due to the loose restrictions on the source data. Howeve…
ALT transforms time series data for better classification.
problem Efficiently classifying time series data with varying temporal scales.
method ALT algorithm using variable-length shifted time windows.
result State-of-the-art performance with minimal computational overhead.
For many years, a combination of principal component analysis (PCA) and independent component analysis (ICA) has been used for blind source separation (BSS). However, it remains unclear why these linear methods work well with real-world data that involve nonlinear source mixtures. This work theoretically validates that…