MoE dispatch pattern according to the network topology.

Retrieval-augmented generation (RAG) enhances the reliability of large language model (LLM) answers by integrating external knowledge. However, RAG increases the end-to-end inference time since looking for relevant documents from large vector databases is computationally expensive. To address this, we introduce Proximity, an approximate key-value cache that optimizes the RAG workflow by leveraging similarities in user queries. Instead of treating each query independently, Proximity reuses previously retrieved documents when similar queries appear, reducing reliance on expensive vector database lookups. We evaluate Proximity on the MMLU and MedRAG benchmarks, demonstrating that it significantly improves retrieval efficiency while maintaining response accuracy. Proximity reduces retrieval latency by up to 59% while maintaining accuracy and lowers the computational burden on the vector database. We also experiment with different similarity thresholds and quantify the trade-off between speed and recall. Our work shows that approximate caching is a viable and effective strategy for optimizing RAG-based systems.

Concept drift is among the primary challenges faced by the data stream processing methods. The drift detection strategies, designed to counteract the negative consequences of such changes, often rely on analyzing the problem metafeatures. This work presents the Frequency Filtering Metadescriptor -- a tool for characterizing the data stream that searches for the informative frequency components visible in the sample's feature vector. The frequencies are filtered according to their variance across all available data batches. The presented solution is capable of generating a metadescription of the data stream, separating chunks into groups describing specific concepts on its basis, and visualizing the frequencies in the original spatial domain. The experimental analysis compared the proposed solution with two state-of-the-art strategies and with the PCA baseline in the post-hoc concept identification task. The research is followed by the identification of concepts in the real-world data streams. The generalization in the frequency domain adapted in the proposed solution allows to capture the complex feature dependencies as a reduced number of frequency components, while maintaining the semantic meaning of data.

Deploying Retrieval Augmented Generation (RAG) on resource-constrained edge devices is challenging due to limited memory and processing power. In this work, we propose EdgeRAG which addresses the memory constraint by pruning embeddings within clusters and generating embeddings on-demand during retrieval. To avoid the latency of generating embeddings for large tail clusters, EdgeRAG pre-computes and stores embeddings for these clusters, while adaptively caching remaining embeddings to minimize redundant computations and further optimize latency. The result from BEIR suite shows that EdgeRAG offers significant latency reduction over the baseline IVF index, but with similar generation quality while allowing all of our evaluated datasets to fit into the memory.

The problem of analyzing data streams of very large volumes is important and is very desirable for many application domains. In this paper we present and demonstrate effective working of an algorithm to find clusters and anomalous data points in a streaming datasets. Entropy minimization is used as a criterion for defining and updating clusters formed from a streaming dataset. As the clusters are formed we also identify anomalous datapoints that show up far away from all known clusters. With a number of 2-D datasets we demonstrate the effectiveness of discovering the clusters and also identifying anomalous data points.

Tabular data is considered the last unconquered castle of deep learning, yet the task of data stream classification is stated to be an equally important and demanding research area. Due to the temporal constraints, it is assumed that deep learning methods are not the optimal solution for application in this field. However, excluding the entire -- and prevalent -- group of methods seems rather rash given the progress that has been made in recent years in its development. For this reason, the following paper is the first to present an approach to natural language data stream classification using the sentence space method, which allows for encoding text into the form of a discrete digital signal. This allows the use of convolutional deep networks dedicated to image classification to solve the task of recognizing fake news based on text data. Based on the real-life Fakeddit dataset, the proposed approach was compared with state-of-the-art algorithms for data stream classification based on generalization ability and time complexity.

Streaming data clustering is a popular research topic in the fields of data mining and machine learning. Compared to static data, streaming data, which is usually analyzed in data chunks, is more susceptible to encountering the dynamic cluster imbalanced issue. That is, the imbalanced degree of clusters varies in different streaming data chunks, leading to corruption in either the accuracy or the efficiency of streaming data analysis based on existing clustering methods. Therefore, we propose an efficient approach called Learning Self-Refined Organizing Map (LSROM) to handle the imbalanced streaming data clustering problem, where we propose an advanced SOM for representing the global data distribution. The constructed SOM is first refined for guiding the partition of the dataset to form many micro-clusters to avoid the missing small clusters in imbalanced data. Then an efficient merging of the micro-clusters is conducted through quick retrieval based on the SOM, which can automatically yield a true number of imbalanced clusters. In comparison to existing imbalanced data clustering approaches, LSROM is with a lower time complexity $O(n\log n)$, while achieving very competitive clustering accuracy. Moreover, LSROM is interpretable and insensitive to hyper-parameters. Extensive experiments have verified its efficacy.

The performance of machine learning models often degrades after deployment due to data distribution shifts. In many use cases, it is impossible to calculate the post-deployment performance because labels are unavailable or significantly delayed. Proxy methods for evaluating model performance stability, like drift detection techniques, do not properly quantify data distribution shift impact. As a solution, we propose a robust and accurate performance estimation method for evaluating ML classification models on unlabeled data that accurately quantifies the impact of covariate shift on model performance. We call it multi-calibrated confidence-based performance estimation (M-CBPE). It is model and data-type agnostic and works for any performance metric. It does not require access to the monitored model - it uses the model predictions and probability estimates. M-CBPE does not need user input on the nature of the covariate shift as it fully learns from the data. We evaluate it with over 600 dataset-model pairs from US census data and compare it with multiple benchmarks using several evaluation metrics. Results show that M-CBPE is the best method to estimate the performance of classification models in any evaluation context.

The recent surge in open-source Large Language Models (LLMs), such as LLaMA, Falcon, and Mistral, provides diverse options for AI practitioners and researchers. However, most LLMs have only released partial artifacts, such as the final model weights or inference code, and technical reports increasingly limit their scope to high-level design choices and surface statistics. These choices hinder progress in the field by degrading transparency into the training of LLMs and forcing teams to rediscover many details in the training process. We present LLM360, an initiative to fully open-source LLMs, which advocates for all training code and data, model checkpoints, and intermediate results to be made available to the community. The goal of LLM360 is to support open and collaborative AI research by making the end-to-end LLM training process transparent and reproducible by everyone. As a first step of LLM360, we release two 7B parameter LLMs pre-trained from scratch, Amber and CrystalCoder, including their training code, data, intermediate checkpoints, and analyses (at https://www.llm360.ai). We are committed to continually pushing the boundaries of LLMs through this open-source effort. More large-scale and stronger models are underway and will be released in the future.