arXiv subjects

Dat Pham

Publications and source records attributed to Dat Pham.

5 recordsLinked to original sources

arXiv2026

ByteRover: Agent-Native Memory Through LLM-Curated Hierarchical Context

Memory-Augmented Generation (MAG) extends large language models with external memory to support long-context reasoning, but existing approaches universally treat memory as an external service that agents call into, delegating storage to separate pipelines of chunking, embedding, and graph extraction. This architectural separation means the system that stores knowledge does not understand it, leading to semantic drift between what the agent intended to remember and what the pipeline actually captured, loss of coordination context across agents, and fragile recovery after failures. In this paper, we propose ByteRover, an agent-native memory architecture that inverts the memory pipeline: the same LLM that reasons about a task also curates, structures, and retrieves knowledge. ByteRover represents knowledge in a hierarchical Context Tree, a file-based knowledge graph organized as Domain, Topic, Subtopic, and Entry, where each entry carries explicit relations, provenance, and an Adaptive Knowledge Lifecycle (AKL) with importance scoring, maturity tiers, and recency decay. Retrieval uses a 5-tier progressive strategy that resolves most queries at sub-100 ms latency without LLM calls, escalating to agentic reasoning only for novel questions. Experiments on LoCoMo and LongMemEval demonstrate that ByteRover achieves state-of-the-art accuracy on LoCoMo and competitive results on LongMemEval while requiring zero external infrastructure, no vector database, no graph database, no embedding service, with all knowledge stored as human-readable markdown files on the local filesystem.

cs.AI↗

arXiv2024

Prismatic $F$-gauges and a result of T. Liu

We give a new proof of a recent result of Tong Liu, which gives a general control on the torsion in the graded pieces of the so-called integral Hodge filtration associated to a crystalline Galois lattice. Our approach is stack-theoretic, and is inspired on the one hand by a result of Gee--Kisin on the shape of mod $p$ crystalline Breuil--Kisin modules, and on the other hand by the structures seen on the diffracted Hodge complex studied by Bhatt--Lurie. Along the way, we also obtain an explicit description of the Hodge--Tate locus in the Nygaard stack $\mathcal{O}_K^{\mathcal{N}}$ for a general extension $K/\mathbf{Q}_p$.

math.NT↗

arXiv2023

Prismatic $F$-crystals and $E$-crystalline Galois representations

Let $K$ be a complete discretely valued field of mixed characteristic $(0,p)$ with perfect residue field, and let $E$ be a finite extension of $\mathbf{Q}_p$ contained in $K$. We show that the category of prismatic $F$-crystals on $\mathcal{O}_K$ (relative to $E$ in a suitable sense) is equivalent to the category of $\mathcal{O}_E$-lattices in $E$-crystalline representations defined by Kisin--Ren, extending the main result of \cite{arxiv:2106.14735} in the case $E=\mathbf{Q}_p$. As a key ingredient in the proof, by adapting a lemma of Du--Liu, we prove a general full faithfulness result for certain vector bundles on the prismatic site, which simplifies and refines the key descent step in the approach of Bhatt--Scholze without invoking the Beilinson fibre sequence.

math.NT↗

arXiv2022

The Emerton--Gee stack of rank one $(\varphi,\Gamma)$-modules

We give a classification of rank one $(\varphi,\Gamma)$-modules with coefficients in a $p$-adically complete $\mathbf{Z}_p$-algebra. As a consequence, we obtain a new proof of Proposition 7.2.17 in {arXiv:1908.07185}, which gives an explicit description of the Emerton--Gee stack of $(\varphi,\Gamma)$-modules in the rank one case. In fact, our method also applies in the context of rank one \' etale $\varphi$-modules (i.e. in the absence of a $\Gamma$-action), generalizing another result of Emerton--Gee.

math.NT↗

arXiv2020

Near real-time map building with multi-class image set labelling and classification of road conditions using convolutional neural networks

Weather is an important factor affecting transportation and road safety. In this paper, we leverage state-of-the-art convolutional neural networks in labelling images taken by street and highway cameras located across across North America. Road camera snapshots were used in experiments with multiple deep learning frameworks to classify images by road condition. The training data for these experiments used images labelled as dry, wet, snow/ice, poor, and offline. The experiments tested different configurations of six convolutional neural networks (VGG-16, ResNet50, Xception, InceptionResNetV2, EfficientNet-B0 and EfficientNet-B4) to assess their suitability to this problem. The precision, accuracy, and recall were measured for each framework configuration. In addition, the training sets were varied both in overall size and by size of individual classes. The final training set included 47,000 images labelled using the five aforementioned classes. The EfficientNet-B4 framework was found to be most suitable to this problem, achieving validation accuracy of 90.6%, although EfficientNet-B0 achieved an accuracy of 90.3% with half the execution time. It was observed that VGG-16 with transfer learning proved to be very useful for data acquisition and pseudo-labelling with limited hardware resources, throughout this project. The EfficientNet-B4 framework was then placed into a real-time production environment, where images could be classified in real-time on an ongoing basis. The classified images were then used to construct a map showing real-time road conditions at various camera locations across North America. The choice of these frameworks and our analysis take into account unique requirements of real-time map building functions. A detailed analysis of the process of semi-automated dataset labelling using these frameworks is also presented in this paper.

cs.CV↗