This paper was accepted at WMT conference at EMNLP. The Transformer architecture has two main non-embedding components: Attention and the Feed Forward Network (FFN). Attention captures interdependencies between words regardless of their position, while the FFN non-linearly transforms each input token independently. In this work,… Read More »One Wide Feedforward is All You Need Apple Machine Learning Research
This paper was accepted at the workshop Deep Generative Models for Health at NeurIPS 2023. Cardiovascular diseases (CVDs) are a major global health concern, making the longitudinal monitoring of cardiovascular biomarkers vital for early diagnosis and intervention. A core challenge is the inference of cardiac… Read More »Hybrid Model Learning for Cardiovascular Biomarkers Inference Apple Machine Learning Research
Using a vision-inspired keyword spotting framework, we propose an architecture with input-dependent dynamic depth capable of processing streaming audio. Specifically, we extend a Conformer encoder with trainable binary gates that allow to dynamically skip network modules according to the input audio. Our approach improves detection… Read More »Improving Vision-inspired Keyword Spotting Using a Streaming Conformer Encoder With Input-dependent Dynamic Depth Apple Machine Learning Research
Super-resolution (SR) techniques have recently been proposed to upscale the outputs of neural radiance fields (NeRF) and generate high-quality images with enhanced inference speeds. However, existing NeRF+SR methods increase training overhead by using extra input features, loss functions, and/or expensive training procedures such as knowledge… Read More »FastSR-NeRF: Improving NeRF Efficiency on Consumer Devices with A Simple Super-Resolution Pipeline Apple Machine Learning Research
This paper was accepted at the workshop Machine Learning and the Physical Sciences at NeurIPS 2023. Over the past decades, hemodynamics simulators have steadily evolved and have become tools of choice for studying cardiovascular systems in-silico. This comes naturally at the cost of increasing complexity… Read More »Simulation-based Inference for Cardiovascular Models Apple Machine Learning Research
Two salient limitations have long hindered the relevance of optimal transport methods to machine learning. First, the computational cost of standard sample-based solvers (when used on batches of samples) is prohibitive. Second, the mass conservation constraint makes OT solvers too rigid in practice: because they… Read More »Unbalanced Low-Rank Optimal Transport Solvers Apple Machine Learning Research
This paper was accepted at the workshop on Regulatable ML at NeurIPS 2023. Conformal Prediction (CP) is a method of estimating risk or uncertainty when using Machine Learning to help abide by common Risk Management regulations often seen in fields like healthcare and finance. CP… Read More »Bin Prediction for Better Conformal Prediction Apple Machine Learning Research
One effective method to improve the reasoning skills of LLMs is to employ supervised fine-tuning (SFT) with chain-of-thought (CoT) annotations. However, this approach has limitations in terms of generalization because it heavily depends on the provided CoT data. In scenarios like math problem-solving, each… Read More »ByteDance AI Research Unveils Reinforced Fine-Tuning (ReFT) Method to Enhance the Generalizability of Learning LLMs for Reasoning with Math Problem Solving as an Example Sana Hassan Artificial Intelligence Category – MarkTechPost
Large Language Models (LLMs) are increasingly employed for various domains, with use cases including creative writing, chatbots, and semantic search. Many of these applications are inherently subjective and require generations catering to different demographics, cultural and societal norms, or individual preferences. Through their large-scale… Read More »UCLA Researchers Introduce Group Preference Optimization (GPO): A Machine Learning-based Alignment Framework that Steers Language Models to Preferences of Individual Groups in a Few-Shot Manner Mohammad Arshad Artificial Intelligence Category – MarkTechPost
The inherent capabilities of pretrained large language models are notable, yet achieving desired behaviors often requires additional adaptation. When dealing with models whose weights are kept private, the challenge intensifies, rendering tuning either excessively costly or outright impossible. As a result, striking the right… Read More »Researchers from the University of Washington and Allen Institute for AI Present Proxy-Tuning: An Efficient Alternative to Finetuning Large Language Models Mohammad Asjad Artificial Intelligence Category – MarkTechPost