A Drosophila-connectome-constrained RNN matches EfficientNet/MobileNet navigation performance yet spontaneously retains function under total vision loss and realistic-texture OOD where dropout-trained baselines fail.
hub Canonical reference
Attention is all you need
Canonical reference. 100% of citing Pith papers cite this work as background.
28
Pith papers citing it
Background
100% of classified citations
hub tools
citation-role summary
background 5
citation-polarity summary
roles
background 5polarities
background 5representative citing papers
Pretrained LLMs adapted via convolutional projections and LoRA act as efficient frozen backbones for sensor-based human activity recognition, delivering strong data efficiency and cross-dataset transfer.
A joint fullband-subband model using high-resolution 44.1 kHz audio outperforms standard 16 kHz detectors for singing voice deepfake detection by exploiting spectrum-specific synthesis artifacts.
Extends vLLM with delay-pattern de-interleaving, multi-stream sampling, and co-scheduled CFG to achieve 80% of non-CFG throughput for unified audio tasks while open-sourcing the pipeline.
PhyPush is a physics-guided Transformer that estimates mass and friction from single-push velocity data, claiming over 10% error reduction versus privileged-force baselines in simulation and successful sim-to-real zero-shot transfer.
MDS-DETR introduces a masked duplicate suppressor in self-attention to enable one-to-many supervision inside a single decoder, yielding +2.8 mAP over Deformable-DETR on COCO with 5% more training time and outperforming MR.DETR by 0.3 mAP while training 20% faster.
EVA applies direct model editing to surgically neutralize jailbreak vulnerabilities in LLMs and VLMs by targeting specific neurons while preserving general capabilities.
MSACT improves localization stability and task success rates in limited-data bimanual manipulation by extracting stable 2D attention points and aligning predicted attention sequences across frames without keypoint labels.
A stereo multistage spatial attention deep predictive learning system improves robustness and success rates for real-time mobile manipulation under visual scale variation and disturbances.
Diffusion models for in-context meta-learning of robot dynamics outperform deterministic Transformers in robustness to distribution shifts while enabling real-time operation via warm-started sampling.
Unsupervised contrastive learning with multi-domain equivalent transformations produces robust radio signal embeddings that outperform baselines in few-shot and cross-domain settings.
CFM uses unstable predictions via contrastive learning to improve SST quality on 3 decision policies and 8 languages in MuST-C v1.0.
SPLC uses social preference feedback to auto-generate preference data for offline RL, improving socially compliant crowd robot navigation over baselines.
A generalizable orientation-invariant framework for eating episode detection from wrist accelerometers with an acc2gyro module for acc-only settings, externally validated on CAD, Capture-24, and Physio-ED datasets with F1 scores of 0.751, 0.592, and 0.793.
MASt3R-Nav proposes pixel-relative connectivity graphs from image sequences to create WayPixel Costmaps that condition a controller for improved visual navigation without global geometric consistency.
MTA-RL predicts 3D driving affordances from multi-modal sensors with a transformer and uses them as the observation space for an RL policy, yielding better route completion and generalization than baselines in CARLA urban scenarios.
BudgetFormer dynamically allocates a variable number of attention heads per input via a learned budget and relevance scoring, reducing inference cost on text classification while maintaining accuracy.
REINA-SAN and REINA-TAN add temporal context to information-based read/write policies, improving the quality-latency tradeoff in simultaneous speech translation by up to 7.1% on Normalized Streaming Efficiency.
A rectified flow model trained on 30 actuation-space demonstrations produces control sequences that yield 97.5% grasp success across the workspace, with generalization to object size changes of ±33% and execution speed scaling from 20% to 200%.
MSDformer introduces a multi-scale discrete transformer that tokenizes time series at multiple scales and models them autoregressively in discrete space, claiming superior performance over prior DTM methods with rate-distortion theoretical support.
ADM-Fusion proposes an end-to-end adaptive multi-sensor fusion network using mixture-of-experts routing and cross-task attention for robust ego-motion estimation, trained on simulation then fine-tuned on real data.
A conditional flow matching model generates realistic safety-critical traffic scenarios by turning nominal scenes into dangerous rollouts using combined simulation and real data.
An agentic LLM/LVM framework generates adaptive behavior trees on-the-fly for AV navigation in CARLA+Nav2 simulation, succeeding in obstacle avoidance where static BTs fail.
A hybrid imitation-plus-reinforcement diffusion planner generates more diverse multi-mode trajectories for end-to-end autonomous driving, with a new diversity metric used for evaluation.
citing papers explorer
-
FLYNN: Robust Neural Network for Robot Navigation using Fly Brain Topology
A Drosophila-connectome-constrained RNN matches EfficientNet/MobileNet navigation performance yet spontaneously retains function under total vision loss and realistic-texture OOD where dropout-trained baselines fail.
-
Efficient and Adaptive Human Activity Recognition via LLM Backbones
Pretrained LLMs adapted via convolutional projections and LoRA act as efficient frozen backbones for sensor-based human activity recognition, delivering strong data efficiency and cross-dataset transfer.
-
Joint Fullband-Subband Modeling for High-Resolution SingFake Detection
A joint fullband-subband model using high-resolution 44.1 kHz audio outperforms standard 16 kHz detectors for singing voice deepfake detection by exploiting spectrum-specific synthesis artifacts.
-
An Efficient vLLM-Based Inference Pipeline for Unified Audio Understanding and Generation
Extends vLLM with delay-pattern de-interleaving, multi-stream sampling, and co-scheduled CFG to achieve 80% of non-CFG throughput for unified audio tasks while open-sourcing the pipeline.
-
PhyPush: One Push is All You Need for Sensorless Physical Property Estimation with Physics-Guided Transformers
PhyPush is a physics-guided Transformer that estimates mass and friction from single-push velocity data, claiming over 10% error reduction versus privileged-force baselines in simulation and successful sim-to-real zero-shot transfer.
-
MDS-DETR: DETR with Masked Duplicate Suppressor
MDS-DETR introduces a masked duplicate suppressor in self-attention to enable one-to-many supervision inside a single decoder, yielding +2.8 mAP over Deformable-DETR on COCO with 5% more training time and outperforming MR.DETR by 0.3 mAP while training 20% faster.
-
EVA: Editing for Versatile Alignment against Jailbreaks
EVA applies direct model editing to surgically neutralize jailbreak vulnerabilities in LLMs and VLMs by targeting specific neurons while preserving general capabilities.
-
MSACT: Multistage Spatial Alignment for Stable Low-Latency Fine Manipulation
MSACT improves localization stability and task success rates in limited-data bimanual manipulation by extracting stable 2D attention points and aligning predicted attention sequences across frames without keypoint labels.
-
Stereo Multistage Spatial Attention for Real-Time Mobile Manipulation Under Visual Scale Variation and Disturbances
A stereo multistage spatial attention deep predictive learning system improves robustness and success rates for real-time mobile manipulation under visual scale variation and disturbances.
-
Diffusion Sequence Models for Generative In-Context Meta-Learning of Robot Dynamics
Diffusion models for in-context meta-learning of robot dynamics outperform deterministic Transformers in robustness to distribution shifts while enabling real-time operation via warm-started sampling.
-
Unsupervised Equivalent Contrastive Learning for Radio Signal Recognition
Unsupervised contrastive learning with multi-domain equivalent transformations produces robust radio signal embeddings that outperform baselines in few-shot and cross-domain settings.
-
Contrastive Feedback Mechanism for Simultaneous Speech Translation
CFM uses unstable predictions via contrastive learning to improve SST quality on 3 decision policies and 8 languages in MuST-C v1.0.
-
SPLC: Social Preference Learning for Crowd Robot Navigation
SPLC uses social preference feedback to auto-generate preference data for offline RL, improving socially compliant crowd robot navigation over baselines.
-
Generalizable framework of eating episode detection on free-living wrist-worn wearable data
A generalizable orientation-invariant framework for eating episode detection from wrist accelerometers with an acc2gyro module for acc-only settings, externally validated on CAD, Capture-24, and Physio-ED datasets with F1 scores of 0.751, 0.592, and 0.793.
-
MASt3R-Nav: WayPixel Navigation in Relative 3D Maps
MASt3R-Nav proposes pixel-relative connectivity graphs from image sequences to create WayPixel Costmaps that condition a controller for improved visual navigation without global geometric consistency.
-
MTA-RL: Robust Urban Driving via Multi-modal Transformer-based 3D Affordances and Reinforcement Learning
MTA-RL predicts 3D driving affordances from multi-modal sensors with a transformer and uses them as the observation space for an RL policy, yielding better route completion and generalization than baselines in CARLA urban scenarios.
-
Adaptive Head Budgeting for Efficient Multi-Head Attention
BudgetFormer dynamically allocates a variable number of attention heads per input via a learned budget and relevance scoring, reducing inference cost on text classification while maintaining accuracy.
-
Regularized Entropy Information Adaptation with Temporal-Awareness Networks for Simultaneous Speech Translation
REINA-SAN and REINA-TAN add temporal context to information-based read/write policies, improving the quality-latency tradeoff in simultaneous speech translation by up to 7.1% on Normalized Streaming Efficiency.
-
Lightweight Learning from Actuation-Space Demonstrations via Flow Matching for Whole-Body Soft Robotic Grasping
A rectified flow model trained on 30 actuation-space demonstrations produces control sequences that yield 97.5% grasp success across the workspace, with generalization to object size changes of ±33% and execution speed scaling from 20% to 200%.
-
MSDformer: Multi-scale Discrete Transformer For Time Series Generation
MSDformer introduces a multi-scale discrete transformer that tokenizes time series at multiple scales and models them autoregressively in discrete space, claiming superior performance over prior DTM methods with rate-distortion theoretical support.
-
ADM-Fusion: Adaptive Deep Multi-Sensor Fusion for Robust Ego-Motion Estimation in Diverse Conditions
ADM-Fusion proposes an end-to-end adaptive multi-sensor fusion network using mixture-of-experts routing and cross-task attention for robust ego-motion estimation, trained on simulation then fine-tuned on real data.
-
Conditional Flow-VAE for Safety-Critical Traffic Scenario Generation
A conditional flow matching model generates realistic safety-critical traffic scenarios by turning nominal scenes into dangerous rollouts using combined simulation and real data.
-
From Prompts to Pavement: LMMs-based Agentic Behavior-Tree Generation Framework for Autonomous Vehicles
An agentic LLM/LVM framework generates adaptive behavior trees on-the-fly for AV navigation in CARLA+Nav2 simulation, succeeding in obstacle avoidance where static BTs fail.
-
DIVER: Reinforced Diffusion Breaks Imitation Bottlenecks in End-to-End Autonomous Driving
A hybrid imitation-plus-reinforcement diffusion planner generates more diverse multi-mode trajectories for end-to-end autonomous driving, with a new diversity metric used for evaluation.
-
A Comprehensive Survey on Network Traffic Synthesis: From Statistical Models to Deep Learning
A survey reviewing statistical and deep learning approaches to synthetic network traffic generation, with comparisons, an AI comparison tool, open challenges, and future directions.
-
Sustainable Code Generation Using Large Language Models: A Systematic Literature Review
A systematic review finds research on the sustainability of LLM-generated code to be limited, fragmented, and without accepted frameworks for measurement or benchmarking.
-
Redefining End-of-Life: Intelligent Automation for Electronics Remanufacturing Systems
A literature review of intelligent automation approaches using robotics, AI, and control for disassembly, inspection, sorting, and reprocessing of end-of-life electronics.
- Progressive Semantic Communication for Efficient Edge-Cloud Vision-Language Models