Elastic Spectral State Space Models for Train-Once Budgeted Inference
cs.LG
Submitted: 2026-01-30
Updated: 2026-09-16
Comments: Minor update: added code repository link
Code: https://github.com/songdc98/ES-S
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- Spectral State Space Models
- Distilling the Knowledge in a Neural Network
- On the Opportunities and Risks of Foundation Models
- Once-for-All: Train One Network and Specialize it for Efficient Deployment
- Transformers are SSMs: Generalized Models and Efficient Algorithms Through Structured State Space Duality
- STAT: Shrinking Transformers After Training
- LazyLLM: Dynamic Token Pruning for Efficient Long Context LLM Inference
- Adaptive Computation Time for Recurrent Neural Networks
- A Mixture of $h-1$ Heads is Better than $h$ Heads
- DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter
- Outrageously Large Neural Networks: The Sparsely-Gated Mixture-of-Experts Layer
- Efficiently Modeling Long Sequences with Structured State Spaces
- Simplified State Space Layers for Sequence Modeling
- Reconstructing Brain Causal Dynamics for Subject and Task Fingerprints using fMRI Time-series Data
- Analyzing Multi-Head Self-Attention: Specialized Heads Do the Heavy Lifting, the Rest Can Be Pruned
- A Survey on Knowledge Distillation of Large Language Models
- Slimmable Neural Networks
Related papers
- Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
- AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
- Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
- Convergence issues in Relational Concept Analysis based on AOC-posets
- Beliefs Beyond Posteriors: Local-Consistency Optimisation for Bayesian Neural Networks
- Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks