-
This AI Research from China Introduces 4K4D: A 4D Point Cloud Representation that Supports Hardware Rasterization and Enables Unprecedented Rendering Speed
Dynamic view synthesis is a technique used in computer vision and graphics to reconstruct dynamic 3D scenes from videos. Traditional methods have limitations in terms of rendering speed and quality. However, a new approach called 4K4D has been introduced, which utilizes a 4D point cloud representation and a hybrid appearance model to achieve faster rendering…
-
This AI Paper Introduces Learning from Mistakes (LeMa): Enhancing Mathematical Reasoning in Large Language Models through Error-Driven Learning
A team of researchers from Jiaotong University, Peking University, and Microsoft have developed a method called LeMa that improves the mathematical reasoning abilities of large language models (LLMs) by teaching them to learn from mistakes. They fine-tune the LLMs using mistake-correction data pairs generated by GPT-4. LeMa consistently improves performance across various LLMs and tasks,…
-
Improved DDIM Sampling with Moment Matching Gaussian Mixtures
In this research, a Gaussian Mixture Model (GMM) is proposed as a reverse transition operator in the Denoising Diffusion Implicit Models (DDIM) framework. By constraining the GMM parameters to match the first and second order central moments of the forward marginals, samples of equal or better quality than the original DDIM with Gaussian kernels can…
-
ReLU Strikes Back: Exploiting Activation Sparsity in Large Language Models
Large Language Models (LLMs) with billions of parameters have revolutionized AI but are computationally intensive. This study supports the use of ReLU activation in LLMs as it minimally affects performance but reduces computation and weight transfer. Alternative activation functions like GELU or SiLU are popular but more computationally demanding.
-
Flexible Keyword Spotting based on Homogeneous Audio-Text Embedding
This work proposes a novel architecture to detect user-defined flexible keywords in real-time. The approach involves constructing acoustic embeddings of keywords using graphene-to-phone conversion, and converting phone-to-embedding by looking up the embedding dictionary built during training. The key benefit is the incorporation of both text and audio embedding.
-
Automating Behavioral Testing in Machine Translation
Behavioral testing in NLP evaluates system capabilities by analyzing input-output behavior. However, current tests for Machine Translation are limited and manually created. To overcome this, our proposal suggests using Large Language Models (LLMs) to generate diverse source sentences for testing MT model behavior in various scenarios. Verification ensures expected performance.
-
How to Scale Your EMA
Preserving training dynamics across batch sizes is important for practical machine learning. One tool for achieving this is scaling the learning rate linearly with the batch size. Another tool is the use of model EMA, which creates a functional copy of a target model that gradually moves towards the parameters of the target model using…
-
Diffusion Models as Masked Audio-Video Learners
Recently, a paper on the use of audio-visual synchronization for learning audio-visual representations was accepted at the Machine Learning for Audio Workshop at NeurIPS 2023. The paper discusses the effectiveness of unsupervised training frameworks, particularly the Masked Audio-Video Learners (MAViL) framework, which combines contrastive learning with masked autoencoding.
-
Agnostically Learning Single-Index Models using Omnipredictors
This text introduces a new approach to agnostically learning Single-Index Models (SIMs) with arbitrary monotone and Lipschitz activations. Unlike previous methods, it does not rely on predetermined settings or knowledge of the activation function. Additionally, it only requires the marginal to have bounded second moments, instead of stronger distributional assumptions. The algorithm is based on…
-
PLANNER: Generating Diversified Paragraph via Latent Language Diffusion Model
Autoregressive models for text generation often produce repetitive and low-quality output due to errors accumulating during generation. Exposure bias, the difference between training and inference, is blamed for this. Denoising diffusion models offer an alternative by allowing a model to revise its output, but they are computationally expensive and less fluent for longer text.