Transformer-based Models
PulseAugur coverage of Transformer-based Models — every cluster mentioning Transformer-based Models across labs, papers, and developer communities, ranked by signal.
6 day(s) with sentiment data
-
ProPINN architecture tackles propagation failures in physics-informed neural networks
Researchers have introduced ProPINN, a novel architecture designed to address propagation failures in physics-informed neural networks (PINNs). These failures occur when supervision signals from initial or boundary cond…
-
MARBERT learns emoji pragmatics in Arabic digital discourse
A new study published on arXiv explores how Transformer-based models, specifically MARBERT, can learn interpersonal pragmatics in Arabic digital discourse through the use of emojis. Researchers collected and annotated a…
-
New Benchmark Compares Deep Learning Models for Brain Tumor Segmentation
Researchers have developed a unified benchmark to compare deep learning models for 3D brain tumor segmentation from MRI scans. The study evaluates five state-of-the-art models, including CNNs, Transformer-based models, …
-
New UEP codec slashes AI inference memory costs by up to 62.5%
Researchers have developed a new method for protecting memory in AI inference by analyzing bit-position fault sensitivity in various models and floating-point formats. They found that certain lower-order bits have minim…
-
M2I2HA network advances multi-modal object detection using hypergraph theory
Researchers have introduced M2I2HA, a novel multi-modal object detection network that utilizes hypergraph theory to improve feature extraction and cross-modal alignment. This approach addresses limitations in existing m…
-
New GATAS method generates adversarial inputs for ASR systems
Researchers have developed a novel black-box testing method called GATAS for automated speech recognition (ASR) systems. This approach generates adversarial inputs by manipulating the latent space of a text-to-speech mo…
-
New X-FEMR approach enhances explainability for electronic health record AI models
Researchers have developed X-FEMR, a novel token-level explainability approach for Foundation Models in Electronic Health Records (FEMRs). These models, while effective for clinical prediction tasks, often function as b…
-
New Threshold Gating Primitive Reimagines Neural Network Nonlinearity
Researchers have proposed a new primitive called Threshold Gating (TG) that can achieve neural nonlinearity, a function traditionally handled by activation functions. This TG primitive is shown to be equivalent to stand…
-
New methods enhance unsupervised cross-modal retrieval with limited data · 4 sources tracked
Researchers are developing new methods for unsupervised cross-modal retrieval, aiming to improve efficiency and reduce reliance on large, manually annotated datasets. Papers propose techniques like Attribute-Prompted Ke…
-
Research paper questions LLM pre-training costs for genomics tasks
A new research paper assesses the effectiveness of pre-training large language models (LLMs) for genomics tasks. The study questions whether the significant computational cost of pre-training transformer-based models li…
-
Time Series Models Evaluated for US Influenza Forecasting
A new research paper evaluates various time series forecasting models for predicting seasonal influenza in the United States. The study found that a mixture-of-experts model, which combines multiple pretrained forecaste…
-
AI models' attention topologies mapped to human brain networks
Researchers have developed a novel method to compare the organizational properties of transformer-based AI models by mapping their attention topologies to human brain networks. This approach allows for modality-agnostic…
-
Legal-specific AI models outperform generalist ones in contract classification
A new study published on arXiv evaluates the performance of transformer-based models specifically customized for legal tasks against generalist models in classifying legal contracts. The research found that legal-specif…
-
LeapTS framework reframes time series forecasting as adaptive scheduling
Researchers have introduced LeapTS, a new framework that reframes time series forecasting as an adaptive scheduling problem. This approach moves away from fixed mappings to a dynamic process where a hierarchical control…
-
AI models learn physics of motion-to-radar spectrograms, study finds
Researchers have developed a new framework to assess whether data-driven models that convert motion capture data to radar spectrograms are learning the underlying physics. This framework uses two metrics to measure the …
-
Directed Social Regard: Surfacing Targeted Advocacy, Opposition, Aid, Harms, and Victimization in Online Media
Researchers have developed a new approach called Directed Social Regard (DSR) to analyze sentiment in online text. Unlike traditional sentiment analysis tools that provide a single positive, neutral, or negative score, …
-
Indonesian students show positive sentiment towards AI in higher education
A new study analyzed Indonesian student sentiment regarding AI adoption in higher education, comparing traditional machine learning with Transformer-based deep learning models. The research utilized a dataset of 2,295 l…