Transformer adapts to graphs with adaptive attention and auto-regressive decoding.
problem Transformers struggle with graph data due to non-sequential nature.
method Proposes GRAT, a Transformer variant with adaptive attention and auto-regressive decoding.
result GRAT achieves state-of-the-art performance on molecule property predictions and generation tasks.
This work proposes an efficient autoregressive model for text generation.
problem The challenge of generating high-quality text with autoregressive models.
method Introduces a cascaded decoding approach using Markov transformers to achieve sub-linear parallel time generation.
result Shows competitive accuracy/speed tradeoff compared to existing methods on five machine translation datasets.
Accelerates TPP sampling with speculative decoding for faster sequence generation.
problem Efficiently sampling from complex temporal point processes.
method Adapting speculative decoding techniques from language models to TPPs.
result Achieves significant speedup (2-6x) while maintaining distributional accuracy.
Proposes a faster Transformer decoding method by truncating target-side self-attention windows.
problem Efficiency in Transformer decoding with minimal BLEU score loss.
method N-gram assumption to truncate target-side self-attention windows.
result N-gram masked self-attention model maintains BLEU score for N values from 4 to 8. Unified framework for learning function representations using INRs and Transformers.
problem Scalability and efficiency limitations in existing generative models.
method Integrates INRs and Transformer-based hypernetworks into latent variable models.
result Improved scalability, expressiveness, and generalization over existing models.
Generative AI decodes quantum codes without labeled data.
problem Efficient decoding of quantum error-correcting codes.
method Generative Transformers learn logical operators from unsupervised syndromes.
result Significantly better decoding accuracy than traditional methods.
Streaming ASR with transformer achieves low WER.
problem Real-time ASR with speech recognition.
method Time-restricted self-attention and triggered attention mechanisms.
result 2.8% and 7.2% WER for LibriSpeech test data.
Deep learning aids ADMM-based decoding for binary linear codes.
problem Improving decoding efficiency for binary linear codes.
method Designing a decoding network based on ADMM and deep learning.
result Numerical results show improved performance compared to original ADMM.
ERP improves drug discovery by balancing molecule generation quality and efficiency.
problem Generating valid and optimal molecules from large language models.
method Entropy-Reinforced Planning (ERP) for Transformer Decoding.
result ERP outperforms current state-of-the-art algorithms by 1-5 percent on SARS-CoV-2 and human cancer cell targets.
This paper analyzes speculative decoding, a method to speed up large language model inferences.
problem Theoretical understanding of speculative decoding is lacking.
method Conceptualizes speculative decoding as a markov chain problem and studies its key properties.
result Reveals fundamental connections between LLM components and their impact on decoding efficiency.
We present the Insertion Transformer, an iterative, partially autoregressive model for sequence generation based on insertion operations. Unlike typical autoregressive models which rely on a fixed, often left-to-right ordering of the output, our approach accommodates arbitrary orderings by allowing for tokens to be ins…
Predicts next item in sequential bundles using Transformers.
problem Predicting next item in sequentially consumed bundles.
method Used custom Transformers, GPT-3, LSTM, reinforcement learning, Markov models.
result Custom Transformer with decoder-only architecture most accurate.
OLinear forecasts time series more efficiently by transforming data orthogonally.
problem Efficiently forecasting time series with entangled dependencies.
method OLinear uses OrthoTrans to transform data orthogonally, then applies NormLin for linear layer.
result OLinear achieves state-of-the-art performance with high efficiency.
Few-shot domain adaptation improves autoencoder performance in changing wireless channels.
problem Frequent retraining of autoencoder for low decoding error rate in changing channel conditions is impractical.
method Uses Gaussian mixture density network and class and component-conditional affine transformations for few-shot adaptation.
result Effective adaptation using very small number of target domain samples, improving performance in real mmWave setups.
TIMeSynC combines financial service interactions for intent prediction.
problem Aligning and learning from multi-domain, multi-resolution sequences for accurate intent prediction.
method An encoder-decoder transformer model addressing sequence alignment, temporal dynamics, and dynamic/static sequence combination.
result Significant improvement in intent prediction over existing methods.
A framework to explain decoder-only sequence classification models using intermediate predictions.
problem Explaining predictions of decoder-only sequence classification models.
method Progressive Inference framework with Single Pass-Progressive Inference and Multi Pass-Progressive Inference methods.
result Significantly better attributions compared to prior work on text classification tasks.
The key idea of current deep learning methods for dense prediction is to apply a model on a regular patch centered on each pixel to make pixel-wise predictions. These methods are limited in the sense that the patches are determined by network architecture instead of learned from data. In this work, we propose the dense…
Recently very deep transformers have outperformed conventional bi-directional long short-term memory networks by a large margin in speech recognition. However, to put it into production usage, inference computation cost is still a serious concern in real scenarios. In this paper, we study two different non-autoregressi…
A transformer model improves spell correction with hierarchical attention.
problem Improving spell correction accuracy and speed.
method Multi encoder-single decoder transformer architecture with hierarchical attention.
result Significant improvement in CER, WER, and SER error rates.
New study finds best language model architecture and pretraining objective for zero-shot tasks.
problem Evaluating which language model architectures and pretraining objectives best enable zero-shot generalization.
method Compared three model architectures and two pretraining objectives across 170 billion tokens, with and without finetuning.
result Causal decoder-only models trained on autoregressive language modeling exhibit strongest zero-shot generalization.
A decoder helps interpret neural network layers.
problem Understanding the output of neural network layers.
method Developed a Classifier-Decoder (ClaDec) architecture.
result Reconstructed images from ClaDec are more relevant for classification.
Generative networks have made it possible to generate meaningful signals such as images and texts from simple noise. Recently, generative methods based on GAN and VAE were developed for graphs and graph signals. However, the mathematical properties of these methods are unclear, and training good generative models is di…
New measures link neural representation geometry to decoding ability.
problem Understanding how neural representations relate to decoding ability.
method Showed that popular similarity measures can be interpreted from a decoding perspective.
result Proved that measures like CKA and CCA quantify alignment between optimal linear readouts.
Most machine translation systems generate text autoregressively from left to right. We, instead, use a masked language modeling objective to train a model to predict any subset of the target words, conditioned on both the input text and a partially masked target translation. This approach allows for efficient iterative…
Parallelizes autoregressive generation using VSSM.
problem Autoregressive models' inability to parallelize generation.
method Variational SSM (VSSM) with parallelizable sampling and decoding.
result Parallel generation possible with VSSM.
Additive decoders tackle latent variables and image generation.
problem Latent variables identification and out-of-support image generation in representation learning.
method Additive decoders that can identify latent variables up to permutation and block-wise invertible transformations, and generate novel images by recombining observed factors.
result Additive decoders provide a new setting for nonlinear independent component analysis and can generate novel images by recombining observed factors.
We describe an approach to Grammatical Error Correction (GEC) that is effective at making use of models trained on large amounts of weakly supervised bitext. We train the Transformer sequence-to-sequence model on 4B tokens of Wikipedia revisions and employ an iterative decoding strategy that is tailored to the loosely-…
Data augmentation (DA) is fundamental against overfitting in large convolutional neural networks, especially with a limited training dataset. In images, DA is usually based on heuristic transformations, like geometric or color transformations. Instead of using predefined transformations, our work learns data augmentati…
Objective: A variety of pattern analysis techniques for model training in brain interfaces exploit neural feature dimensionality reduction based on feature ranking and selection heuristics. In the light of broad evidence demonstrating the potential sub-optimality of ranking based feature selection by any criterion, we …
Geometry-aware models improve cross-subject EEG decoding accuracy.
problem Strong inter-subject variability in motor imagery decoding.
method Discriminative Congruence Transform (DCT), Deep Linear DCT (DLDCT), Deep DCT-UNet (DDCT-UNet).
result Improves transductive cross-subject accuracy by 2-3%.
Steganography is the science of hiding a secret message within an ordinary public message, which is referred to as Carrier. Traditionally, digital signal processing techniques, such as least significant bit encoding, were used for hiding messages. In this paper, we explore the use of deep neural networks as steganograp…
Enhanced transformer converts whispered speech to natural speech.
problem Machine recognition of whispered speech is challenging.
method Proposes an enhanced transformer architecture trained end-to-end using supervised learning.
result Similar formant distributions of converted speech to groundtruth.
VED framework learns low-dimensional latent representations of physical systems.
problem Learning latent representations of complex physical systems.
method Variational Encoder-Decoder (VED) framework with KL divergence and covariance regularization.
result VED achieves lower-dimensional latent representations with improved feature disentanglement.
While deep learning has received a surge of interest in a variety of fields in recent years, major deep learning models barely use complex numbers. However, speech, signal and audio data are naturally complex-valued after Fourier Transform, and studies have shown a potentially richer representation of complex nets. In …
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.
VAIOM models financial returns using continuous input and categorical output.
problem Modeling continuous, noisy, and heterogeneous financial data.
method VAIOM is a decoder-only Transformer that separates input representation from output likelihood.
result VAIOM models outperform fixed single-bar LightGBM baseline in both Test halves.
Paper improves text-to-SQL models with schema-aware denoising.
problem Text-to-SQL models struggle with schema linking and grammar correctness.
method Adapts transformer-based seq-to-seq model with SeaD denoising objectives and clause-sensitive decoding.
result Improves seq-to-seq model performance on WikiSQL benchmark.
Recent progress on deep learning has made it possible to automatically transform the screenshot of Graphic User Interface (GUI) into code by using the encoder-decoder framework. While the commonly adopted image encoder (e.g., CNN network), might be capable of extracting image features to the desired level, interpreting…
Transformation Equivariant Representations (TERs) aim to capture the intrinsic visual structures that equivary to various transformations by expanding the notion of {\em translation} equivariance underlying the success of Convolutional Neural Networks (CNNs). For this purpose, we present both deterministic AutoEncoding…
The paper accelerates LLM inference by adding early exit heads trained in a self-supervised manner.
problem Inference speed in large language models (LLMs) is slow and resource-intensive.
method Adding self-supervised early exit heads at intermediate transformer layers to stop computation early based on confidence thresholds.
result Entropy provides the most reliable confidence metric for stopping computation early.
Image compression is an essential approach for decreasing the size in bytes of the image without deteriorating the quality of it. Typically, classic algorithms are used but recently deep-learning has been successfully applied. In this work, is presented a deep super-resolution work-flow for image compression that maps …
C3 compresses images and videos with low complexity and high performance.
problem High complexity and low performance in neural compression models.
method Overfits a small model to each image or video separately, improving RD performance with low complexity.
result Matches the RD performance of state-of-the-art neural and video codecs with significantly lower decoding complexity.
In many state-of-the-art compression systems, signal transformation is an integral part of the encoding and decoding process, where transforms provide compact representations for the signals of interest. This paper introduces a class of transforms called graph-based transforms (GBTs) for video compression, and proposes…
Here, we present a novel approach to solve the problem of reconstructing perceived stimuli from brain responses by combining probabilistic inference with deep learning. Our approach first inverts the linear transformation from latent features to brain responses with maximum a posteriori estimation and then inverts the …
DOS improves language model generation by considering inter-token dependencies.
problem Lack of sequence-level information and inter-token dependencies in existing decoding strategies.
method Dependency-Oriented Sampler (DOS) that uses attention matrices to approximate inter-token dependencies.
result DOS consistently achieves superior performance on code generation and mathematical reasoning tasks.
Decoding language representations directly from the brain can enable new Brain-Computer Interfaces (BCI) for high bandwidth human-human and human-machine communication. Clinically, such technologies can restore communication in people with neurological conditions affecting their ability to speak. In this study, we prop…
MQTransformer improves forecast accuracy with context-aware attention.
problem Improving probabilistic demand prediction accuracy.
method Incorporates Transformer architectures for context alignment and feedback-aware attention.
result Significant improvements in forecast accuracy, reducing excess variability.
We develop a deep autoencoder architecture that can be used to find a coordinate transformation which turns a nonlinear PDE into a linear PDE. Our architecture is motivated by the linearizing transformations provided by the Cole-Hopf transform for Burgers equation and the inverse scattering transform for completely int…