Rise
PulseAugur coverage of Rise — every cluster mentioning Rise across labs, papers, and developer communities, ranked by signal.
- 2026-05-18 research_milestone Publication of a new framework called RISE for improving Rhetorical Role Labeling on hard examples. source
-
RISE method enhances language model training via self-extrapolation
Researchers have introduced RISE, a novel method for improving language model post-training through self-extrapolating policy distillation. This technique constructs a synthetic teacher from the model's own reinforcemen…
-
New framework RISE enhances multilingual MoE models by isolating language experts
Researchers have identified a phenomenon called Language Routing Isolation in multilingual Mixture-of-Experts (MoE) models, where different language sets activate distinct sets of experts. This isolation pattern varies …
-
New World Action Models Enhance AI Agents in Games and Robotics
Researchers are developing novel World Action Models (WAMs) to improve agent performance in video games and robotics. GameWAM, for instance, unifies visual prediction and action generation for native game control, demon…
-
New framework TabNSM enhances tabular regression with sparse interaction modeling
Researchers have introduced TabNSM, a new framework designed to tackle large-scale, high-dimensional tabular regression challenges. This framework integrates adaptive sparse interaction modeling for efficient feature pr…
-
New RISE framework integrates 3D tracking and vision-language reasoning for roadside analysis
Researchers have introduced RISE (Roadside Infrastructure Sequence Understanding and Evaluation), a novel framework designed for analyzing roadside traffic sequences. This framework integrates metric 3D tracking using i…
-
New RISE framework enhances low-light images using structural illumination
Researchers have developed a new framework called Relative Illumination Structure Estimation (RISE) for unsupervised low-light image enhancement. This method decouples the relative illumination structure from absolute e…
-
OpenDriveLab's Li Chen proposes compositional world models for safer AI policies
OpenDriveLab's Li Chen presented a compositional world model approach at RSS 2026, separating prediction and evaluation components for embodied AI policies. This decoupling aims to improve safety and inspectability by a…
-
New dataset DisciplineGen-1M boosts AI visual generation for academic content
Researchers have introduced DisciplineGen-1M, a large-scale dataset designed to improve the accuracy of AI models in generating and editing knowledge-intensive visual content. This dataset comprises 1.2 million samples …
-
Radar system RISE enables privacy-preserving indoor scene understanding
Researchers have introduced RISE, a novel system and benchmark for understanding indoor environments using a single, static radar sensor. Unlike optical sensors, radar offers privacy and can penetrate obstacles, but typ…
-
Rust library RISE boosts inverted index search performance
Researchers have developed RISE, a new library in Rust for creating and querying inverted indexes, which are essential for efficient text search. The library aims to provide high performance and safety by utilizing Rust…
-
New research tackles multilingual adaptation in Mixture-of-Experts models
Two new research papers explore the adaptation of Mixture-of-Experts (MoE) models for multilingual tasks. One paper analyzes how language specialization emerges in MoE models during continual pre-training, finding that …
-
New audit protocol assesses AI explanation faithfulness in visual inspection
Researchers have developed a new method for auditing the explanations generated by deep learning models used in industrial visual inspection. This "architecture-aware" protocol assesses how faithfully an explanation met…
-
Swedish researcher awarded for Ukraine defense training and AI insights
A Swedish researcher received a military medal for contributions to psychological defense training and knowledge exchange with Ukraine. The award highlights the importance of rapid learning and adaptation in modern warf…
-
New framework boosts Rhetorical Role Labeling accuracy on hard text examples
Researchers have developed RISE, a new framework designed to improve the accuracy of Rhetorical Role Labeling (RRL) on challenging text segments. This method operates at inference time, semantically reranking prediction…
-
New theory explains why Zeroth-Order adaptation reduces model forgetting
Researchers have developed a new theoretical framework, Randomized Shaping Theory, to explain why Zeroth-Order (ZO) adaptation methods in continual learning may lead to less forgetting than first-order (FO) methods. The…