# AMemOps: Evaluating Memory Updates in Long Audio Streams

arXiv: https://arxiv.org/abs/2610.18652 · HTML: https://mlsanity.com/paper/2610.18652/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 28 of its 31 references that are in the map.
- First posted on arXiv 13 Oct 2026.

## Builds on (most cited first, 28 of 28)

- [RoFormer: Enhanced Transformer with Rotary Position Embedding](https://mlsanity.com/paper/2104.09864/) (2021)
- [Transformers are RNNs: Fast Autoregressive Transformers with Linear Attention](https://mlsanity.com/paper/2006.16236/) (2020)
- [Transformers are SSMs: Generalized Models and Efficient Algorithms Through Structured State Space Duality](https://mlsanity.com/paper/2405.21060/) (2024)
- [Neural Turing Machines](https://mlsanity.com/paper/1410.5401/) (2014)
- [Gated Delta Networks: Improving Mamba2 with Delta Rule](https://mlsanity.com/paper/2412.06464/) (2024)
- [Gated Linear Attention Transformers with Hardware-Efficient Training](https://mlsanity.com/paper/2312.06635/) (2023)
- [Linear Transformers Are Secretly Fast Weight Programmers](https://mlsanity.com/paper/2102.11174/) (2021)
- [Parallelizing Linear Transformers with the Delta Rule over Sequence Length](https://mlsanity.com/paper/2406.06484/) (2024)
- [Query-Key Normalization for Transformers](https://mlsanity.com/paper/2010.04245/) (2020)
- [Neural Audio Synthesis of Musical Notes with WaveNet Autoencoders](https://mlsanity.com/paper/1704.01279/) (2017)
- [Zoology: Measuring and Improving Recall in Efficient Language Models](https://mlsanity.com/paper/2312.04927/) (2023)
- [Kimi Linear: An Expressive, Efficient Attention Architecture](https://mlsanity.com/paper/2510.26692/) (2025)
- [Audio Flamingo 2: An Audio-Language Model with Long-Audio Understanding and Expert Reasoning Abilities](https://mlsanity.com/paper/2503.03983/) (2025)
- [Mechanistic Design and Scaling of Hybrid Architectures](https://mlsanity.com/paper/2403.17844/) (2024)
- [Temporal Reasoning via Audio Question Answering](https://mlsanity.com/paper/1911.09655/) (2019)
- [Gated DeltaNet-2: Decoupling Erase and Write in Linear Attention](https://mlsanity.com/paper/2605.22791/) (2026)
- [AudioMarathon: A Comprehensive Benchmark for Long-Context Audio Understanding and Efficiency in Audio LLMs](https://mlsanity.com/paper/2510.07293/) (2025)
- [Audio MultiChallenge: A Multi-Turn Evaluation of Spoken Dialogue Systems on Natural Human Interaction](https://mlsanity.com/paper/2512.14865/) (2025)
- [Mechanistic evaluation of Transformers and state space models](https://mlsanity.com/paper/2505.15105/) (2025)
- [MemOps: Benchmarking Lifecycle Memory Operations in Long-Horizon Conversations](https://mlsanity.com/paper/2607.12893/) (2026)
- [Erase-then-Delta Attention: Decoupling Erase and Write Addresses in Delta-Rule Linear Attention](https://mlsanity.com/paper/2606.26560/) (2026)
- [NAAQA: A Neural Architecture for Acoustic Question Answering](https://mlsanity.com/paper/2106.06147/) (2021)
- [Kernelized Linear Attention: Breaking the Capacity Wall with Symmetric Cones](https://mlsanity.com/paper/2607.17419/) (2026)
- [Event-Grounded Question Answering over Long Audio via Structured Retrieval](https://mlsanity.com/paper/2602.14612/) (2026)
- [Why Can't They Remember? Uncovering Representation and Retrieval Bottlenecks in Multi-Turn Acoustic Memory](https://mlsanity.com/paper/2605.27039/) (2026)
- [RAIL: Rethinking Auditory Intelligence in Large Audio-Language Models with a CHC-Grounded Benchmark](https://mlsanity.com/paper/2606.11260/) (2026)
- [The Query Knows What to Forget: A Second Erase Direction for Linear Attention](https://mlsanity.com/paper/2608.13668/) (2026)
- [Do SpeechLMs Hear Their Own Opinions? Diagnosing and Mitigating Previous-Belief Contamination in Streaming Emotion Understanding](https://mlsanity.com/paper/2608.20769/) (2026)

## Abstract (arXiv)

Audio memory must recognise recurring sources across acoustic changes while keeping their attributes up to date. Overall accuracy does not reveal how these demands interact. We introduce Audio Memory Operations (AMemOps), a benchmark for latest-attribute retrieval, preservation of unchanged attributes and absent-source rejection in long audio streams. Paired controls vary recordings, training identities and earlier events to examine their effects on updating. Training on other recordings of the test instruments increases a reference delta-rule memory's update advantage over matched attention by 13.4 percentage points. At retrieval delays of 120,s, matched histories separate benefits of consistent repetition from costs of conflicting earlier values, while additive-memory controls identify benefits of erasure.