Copertina

Dynamic Context Adaptation and Information Flow Control in Transformers: Introducing the Evaluator Adjuster Unit and Gated Residual Connections

Fuente: arXiv

Salvato in:

Dettagli Bibliografici
Autore principale:	Dhayalkar, Sahil Rajesh
Natura:	Preprint
Pubblicazione:	2024
Soggetti:	Machine Learning Artificial Intelligence
Accesso online:	Acceder al recurso
Tags:	Aggiungi Tag Nessun Tag, puoi essere il primo ad aggiungerne!!

Documenti analoghi

Attention as Binding: A Vector-Symbolic Perspective on Transformer Reasoning
di: Dhayalkar, Sahil Rajesh
Pubblicazione: (2025)

A Combinatorial Theory of Dropout: Subnetworks, Graph Geometry, and Generalization
di: Dhayalkar, Sahil Rajesh
Pubblicazione: (2025)

Reasoning Stabilization Point: A Training-Time Signal for Stable Evidence and Shortcut Reliance
di: Dhayalkar, Sahil Rajesh
Pubblicazione: (2026)

The Geometry of ReLU Networks through the ReLU Transition Graph
di: Dhayalkar, Sahil Rajesh
Pubblicazione: (2025)

Neural Networks as Universal Finite-State Machines: A Constructive Deterministic Finite Automaton Theory
di: Dhayalkar, Sahil Rajesh
Pubblicazione: (2025)

Logic-Gated Time-Shared Feedforward Networks for Alternating Finite Automata: Exact Simulation and Learnability
di: Dhayalkar, Sahil Rajesh
Pubblicazione: (2026)

Symbolic Feedforward Networks for Probabilistic Finite Automata: Exact Simulation and Learnability
di: Dhayalkar, Sahil Rajesh
Pubblicazione: (2025)

Discrete Functional Geometry of ReLU Networks via ReLU Transition Graphs
di: Dhayalkar, Sahil Rajesh
Pubblicazione: (2025)

A Constructive Framework for Nondeterministic Automata via Time-Shared, Depth-Unrolled Feedforward Networks
di: Dhayalkar, Sahil Rajesh
Pubblicazione: (2025)

Human Activity Recognition using Smartphones
di: Sonawane, Mayur, et al.
Pubblicazione: (2024)

MUDDFormer: Breaking Residual Bottlenecks in Transformers via Multiway Dynamic Dense Connections
di: Xiao, Da, et al.
Pubblicazione: (2025)

Gradient Residual Connections
di: Pan, Yangchen, et al.
Pubblicazione: (2026)

PrismFlow: Residual Dynamics for Flow Matching in Time-Series Generation
di: Zhang, Junru, et al.
Pubblicazione: (2026)

ResidualDroppath: Enhancing Feature Reuse over Residual Connections
di: Park, Sejik
Pubblicazione: (2024)

Minion Gated Recurrent Unit for Continual Learning
di: Zyarah, Abdullah M., et al.
Pubblicazione: (2025)

Neural Additive Experts: Context-Gated Experts for Controllable Model Additivity
di: Xiong, Guangzhi, et al.
Pubblicazione: (2026)

Gate Recurrent Unit for Efficient Industrial Gas Identification
di: Wang, Ding
Pubblicazione: (2024)

GPU Memory Requirement Prediction for Deep Learning Task Based on Bidirectional Gated Recurrent Unit Optimization Transformer
di: Wang, Chao, et al.
Pubblicazione: (2025)

FedGSCA: Medical Federated Learning with Global Sample Selector and Client Adaptive Adjuster under Label Noise
di: Ye, Mengwen, et al.
Pubblicazione: (2025)

Dynamics of the Transformer Residual Stream: Coupling Spectral Geometry to Network Topology
di: Fernando, Jesseba, et al.
Pubblicazione: (2026)

Contrastive Residual Energy Test-time Adaptation
di: Han, Yewon, et al.
Pubblicazione: (2025)

Reinforcement Learning for Synchronised Flow Control in a Dual-Gate Resin Infusion System
di: Camacho-Sánchez, Miguel, et al.
Pubblicazione: (2025)

Gated Adaptation for Continual Learning in Human Activity Recognition
di: Azghan, Reza Rahimi, et al.
Pubblicazione: (2026)

BiTA: Bidirectional Gated Recurrent Unit-Transformer Aggregator in a Temporal Graph Network Framework for Alert Prediction in Computer Networks
di: Nayeri, Zahra Makki, et al.
Pubblicazione: (2026)

No learning rates needed: Introducing SALSA -- Stable Armijo Line Search Adaptation
di: Kenneweg, Philip, et al.
Pubblicazione: (2024)

PETAH: Parameter Efficient Task Adaptation for Hybrid Transformers in a resource-limited Context
di: Augustin, Maximilian, et al.
Pubblicazione: (2024)

Turbo Connection: Reasoning as Information Flow from Higher to Lower Layers
di: Tang, Mohan, et al.
Pubblicazione: (2026)

A Method for Optimizing Connections in Differentiable Logic Gate Networks
di: Mommen, Wout, et al.
Pubblicazione: (2025)

CosineGate: Semantic Dynamic Routing via Cosine Incompatibility in Residual Networks
di: Thota, Yogeswar Reddy
Pubblicazione: (2025)

Ablate and Rescue: A Causal Analysis of Residual Stream Hyper-Connections
di: Peng, William, et al.
Pubblicazione: (2026)

Flash STU: Fast Spectral Transform Units
di: Liu, Y. Isabel, et al.
Pubblicazione: (2024)

Symmetry Reveals Layerwise Dynamics: How Transformers Perform In-Context Classification
di: Lutz, Patrick, et al.
Pubblicazione: (2026)

PaCA: Partial Connection Adaptation for Efficient Fine-Tuning
di: Woo, Sunghyeon, et al.
Pubblicazione: (2025)

Fine-Tuned In-Context Learners for Efficient Adaptation
di: Bornschein, Jorg, et al.
Pubblicazione: (2025)

Context-Aware Self-Adaptation for Domain Generalization
di: Yan, Hao, et al.
Pubblicazione: (2025)

In-Context Reward Adaptation for Robust Preference Modeling
di: Sun, Zhenyu, et al.
Pubblicazione: (2026)

Twin Transformer using Gated Dynamic Learnable Attention mechanism for Fault Detection and Diagnosis in the Tennessee Eastman Process
di: Labbaf-Khaniki, Mohammad Ali, et al.
Pubblicazione: (2024)

The Phasor Transformer: Resolving Attention Bottlenecks on the Unit Circle
di: Sigdel, Dibakar
Pubblicazione: (2026)

How Transformers Learn In-Context Recall Tasks? Optimality, Training Dynamics and Generalization
di: Nguyen, Quan, et al.
Pubblicazione: (2025)

ContextFlow++: Generalist-Specialist Flow-based Generative Models with Mixed-Variable Context Encoding
di: Gudovskiy, Denis, et al.
Pubblicazione: (2024)