AI research from July 2026 Page 5

This paper introduces a memory-efficient training technique called Hierarchical Global Attention to process significantly longer token sequences on constrained GPU hardware.

The HDR model uses a tree-structured hierarchy to balance logical consistency in multi-step visual reasoning with efficient streaming performance.

SUFLECA improves zero-shot CAD-to-image alignment by scaling up geometry-aware feature learning, leading to significantly better accuracy on benchmarks like ScanNet25k.

AeroAct adapts a pretrained video diffusion transformer to connect visual language task specification, predictive action generation, and closed-loop quadrotor execution without generating future video at deployment time.

The researchers demonstrate that current AI coding agents frequently fail to detect malicious package installations when following project setup documentation.

The researchers developed the MM-IssueLoc benchmark to evaluate how visual evidence like screenshots affects AI-driven repository-level issue localization.

The paper demonstrates that decomposing complex queries into smaller attribute-based sub-problems improves an LLM's consistency and alignment with real-world data.

The researchers demonstrate that the bias in specific statistical sampling algorithms can be constrained relative to the system dimension when variables share sparse interactions.

NeuronSoup replaces backpropagation with an asynchronous evolutionary algorithm that uses discrete event simulation to train models for efficient, variable-depth processing.

The paper presents data-driven algorithms to determine the optimal interval for replacing machine parts when their lifetime distribution is unknown.

The paper presents a sentiment classifier that integrates daily Bitcoin on-chain metrics with Twitter sentiment analysis to identify market phases.

WanSong v1.0 is a music generation model that uses a hybrid transformer architecture and a dual-stem output strategy to improve the fidelity and separation of vocals and background music.

The paper introduces a probing technique that detects when LLM-generated instructions pose physical risks despite appearing linguistically safe.

This paper presents an in-place recipe to expand pre-trained model tokenizers to better support underrepresented languages without discarding existing model weights.

The authors introduce a mutable sketch method that updates user preferences in log-time to allow for real-time recommendations without needing to retrain the entire model.

The paper introduces a cost evaluation framework for security agents that measures economic efficiency alongside performance on offensive and defensive benchmarks.

The paper introduces Online Neural Space Time Memory to enable persistent, long-horizon novel view synthesis for streaming video without the computational overhead of per-frame updates.

The researchers developed a model that performs video object removal in a single inference step by eliminating the need for external draft priors.

Project Kaleidoscope introduces a workflow for calibrating automated LLM judges against human labels to evaluate real world AI applications.

The paper introduces a method called SEED that improves agentic performance by having the model analyze its own past trajectories to generate dense feedback for training.