Compressed Concatenation of Small Embedding Models

Embedding models are central to dense retrieval, semantic search, and recommendation systems, but their size often makes them impractical to deploy in resource-constrained environments such as browsers or edge devices. While smaller embedding models offer practical advantages, they typically underperform compared to their larger counterparts. To bridge this gap, we demonstrate that concatenating the raw embedding vectors of multiple small models can outperform a single larger baseline on standard retrieval benchmarks. To overcome the resulting high dimensionality of naive concatenation, we introduce a lightweight unified decoder trained with a Matryoshka Representation Learning (MRL) loss. This decoder maps the high-dimensional joint representation to a low-dimensional space, preserving most of the original performance without fine-tuning the base models. We also show that while concatenating more base models yields diminishing gains, the robustness of the decoder's representation under compression and quantization improves. Our experiments show that, on a subset of MTEB retrieval tasks, our concat-encode-quantize pipeline recovers 89\% of the original performance with a 48x compression factor when the pipeline is applied to a concatenation of four small embedding models.

WebFAQ: A Multilingual Collection of Natural Q&A Datasets for Dense Retrieval

We present WebFAQ, a large-scale collection of open-domain question answering datasets derived from FAQ-style schema.org annotations. In total, the data collection consists of 96 million natural question-answer (QA) pairs across 75 languages, including 47 million (49%) non-English samples. WebFAQ further serves as the foundation for 20 monolingual retrieval benchmarks with a total size of 11.2 million QA pairs (5.9 million non-English). These datasets are carefully curated through refined filtering and near-duplicate detection, yielding high-quality resources for training and evaluating multilingual dense retrieval models. To empirically confirm WebFAQ's efficacy, we use the collected QAs to fine-tune an in-domain pretrained XLM-RoBERTa model. Through this process of dataset-specific fine-tuning, the model achieves significant retrieval performance gains, which generalize - beyond WebFAQ - to other multilingual retrieval benchmarks evaluated in zero-shot setting. Last but not least, we utilize WebFAQ to construct a set of QA-aligned bilingual corpora spanning over 1000 language pairs using state-of-the-art bitext mining and automated LLM-assessed translation evaluation. Due to our advanced, automated method of bitext dataset generation, the resulting bilingual corpora demonstrate higher translation quality compared to similar datasets. WebFAQ and all associated resources are publicly available on GitHub and HuggingFace.

A Survey of Web Content Control for Generative AI

The groundbreaking advancements around generative AI have recently caused a wave of concern culminating in a row of lawsuits, including high-profile actions against Stability AI and OpenAI. This situation of legal uncertainty has sparked a broad discussion on the rights of content creators and publishers to protect their intellectual property on the web. European as well as US law already provides rough guidelines, setting a direction for technical solutions to regulate web data use. In this course, researchers and practitioners have worked on numerous web standards and opt-out formats that empower publishers to keep their data out of the development of generative AI models. The emerging AI/ML opt-out protocols are valuable in regards to data sovereignty, but again, it creates an adverse situation for a site owners who are overwhelmed by the multitude of recent ad hoc standards to consider. In our work, we want to survey the different proposals, ideas and initiatives, and provide a comprehensive legal and technical background in the context of the current discussion on web publishers control.