We present MA-RAG, a Multi-Agent framework for Retrieval-Augmented Generation (RAG) that addresses the inherent ambiguities and reasoning challenges in complex information-seeking tasks. Unlike conventional RAG methods that rely on end-to-end fine-tuning or isolated component enhancements, MA-RAG orchestrates a collaborative set of specialized AI agents: Planner, Step Definer, Extractor, and QA Agents, each responsible for a distinct stage of the RAG pipeline. By decomposing tasks into subtasks such as query disambiguation, evidence extraction, and answer synthesis, and enabling agents to communicate intermediate reasoning via chain-of-thought prompting, MA-RAG progressively refines retrieval and synthesis while maintaining modular interpretability. Extensive experiments on multi-hop and ambiguous QA benchmarks, including NQ, HotpotQA, 2WikimQA, and TriviaQA, demonstrate that MA-RAG significantly outperforms standalone LLMs and existing RAG methods across all model scales. Notably, even a small LLaMA3-8B model equipped with MA-RAG surpasses larger standalone LLMs, while larger variants (LLaMA3-70B and GPT-4o-mini) set new state-of-the-art results on challenging multi-hop datasets. Ablation studies reveal that both the planner and extractor agents are critical for multi-hop reasoning, and that high-capacity models are especially important for the QA agent to synthesize answers effectively. Beyond general-domain QA, MA-RAG generalizes to specialized domains such as medical QA, achieving competitive performance against domain-specific models without any domain-specific fine-tuning. Our results highlight the effectiveness of collaborative, modular reasoning in retrieval-augmented systems: MA-RAG not only improves answer accuracy and robustness but also provides interpretable intermediate reasoning steps, establishing a new paradigm for efficient and reliable multi-agent RAG.

LLMs demonstrate remarkable capabilities in following natural language instructions, largely due to instruction-tuning on high-quality datasets. While synthetic data generation has emerged as a scalable approach for creating such datasets, maintaining consistent quality standards remains challenging. Recent approaches incorporate feedback to improve data quality, but typically operate at the sample level, generating and applying feedback for each response individually. In this work, we propose Reference-Level Feedback, a novel methodology that instead collects feedback based on high-quality reference samples from carefully curated seed data. We use this feedback to capture rich signals of desirable characteristics and propagate it throughout the data synthesis process. We present REFED, a dataset of 10K instruction-response pairs synthesized using such feedback. We demonstrate the effectiveness of our approach by showing that Llama-3.1-8B-Instruct finetuned on REFED achieves state-of-the-art performance among similar-sized SFT-based models on AlpacaEval 2.0 and strong results on Arena-Hard. Through extensive experiments, we show that our approach consistently outperforms traditional sample-level feedback methods with significantly fewer feedback collections and improves performance across different model architectures.

In recent years, artificial intelligence (AI) has made incredible strides in its ability to generate human-like text. As a result, AI writing is becoming increasingly commonplace, with businesses and organisations using it to create everything from marketing copy to financial reports. While AI writing is still in its early stages and far from perfect, it's clear that it poses a threat to the livelihood of professional writers. After all, if a machine can produce text that is indistinguishable from that of a human writer, why would anyone need to hire a real person to do the job? It's not just low-skilled jobs like content writing that are at risk of being automated by AI.

Join us on November 9 to learn how to successfully innovate and achieve efficiency by upskilling and scaling citizen developers at the Low-Code/No-Code Summit. For Dave Rogenmoser, CEO of AI content platform Jasper -- which raised $125 million in funding a week ago -- the sheer level of hype and scale of chatter around generative AI last week was unexpected. Jasper's announcement came just one day after Stability AI, which developed its text-to-image generator Stable Diffusion, announced its own massive $101 million raise. "I didn't know Stability was going to announce on Monday -- and then ours stacking on that definitely hyped up the whole market," he said. But Rogenmoser says that hype aside, generative AI -- which describes artificial intelligence using unsupervised learning algorithms to create new digital images, video, audio, text or code -- is no flash in the pan.

The aim of content marketing is to attract, engage, and retain customers. It takes many forms, including videos, podcasts, graphics, articles, and whitepapers. Each of those could have a sub-task. This article focuses on attracting an audience -- driving top-of-the-funnel prospects -- with blog content. A blog post that ranks well on search engine results pages must include the words and phrases of searchers.

Nonnegative least squares (NNLS) problems arise in models that rely on additive linear combinations. In particular, they are at the core of nonnegative matrix factorization (NMF) algorithms. The nonnegativity constraint is known to naturally favor sparsity, that is, solutions with few non-zero entries. However, it is often useful to further enhance this sparsity, as it improves the interpretability of the results and helps reducing noise. While the $\ell_0$-"norm", equal to the number of non-zeros entries in a vector, is a natural sparsity measure, its combinatorial nature makes it difficult to use in practical optimization schemes. Most existing approaches thus rely either on its convex surrogate, the $\ell_1$-norm, or on heuristics such as greedy algorithms. In the case of multiple right-hand sides NNLS (MNNLS), which are used within NMF algorithms, sparsity is often enforced column- or row-wise, and the fact that the solution is a matrix is not exploited. In this paper, we first introduce a novel formulation for sparse MNNLS, with a matrix-wise $\ell_0$ sparsity constraint. Then, we present a two-step algorithm to tackle this problem. The first step uses a homotopy algorithm to produce the whole regularization path for all the $\ell_1$-penalized NNLS problems arising in MNNLS, that is, to produce a set of solutions representing different tradeoffs between reconstruction error and sparsity. The second step selects solutions among these paths in order to build a sparsity-constrained matrix that minimizes the reconstruction error. We illustrate the advantages of our proposed algorithm for the unmixing of facial and hyperspectral images.

Experienced programmers understand perfectly well that in development they spend most of the time reading code and therefore they treat the process of writing code with the deepest trepidation (and sometimes with fanaticism). To write quality and maintainable code, you need to take the time to write tests and integrate QA tools. There is a whole technique aimed at test-driven development (TDD) and I will not devote this article to the topic of testing as such. Tests are absolutely necessary and there is nothing to discuss. In this article, we are going to talk about tools that help you write quality Python code.

Cisco today introduced machine learning capabilities and tighter integration between service providers and vendors in its IoT management platform. Cisco IoT Control Center now includes machine learning models to identify anomalies and resolve problems before they impact IoT services. The feature also enables service providers to alert customers of errant or otherwise unused devices, and therefore implement greater control over connected devices. Cisco's IoT platform is largely the result of its $1.4 billion acquisition of Jasper in 2016, but the effort has grown considerably during the last four years. Cisco has partnerships and IoT management reselling agreements with 52 network operators and claims to be the No. 1 IoT management platform provider for connected cars.

Choice functions constitute a simple, direct and very general mathematical framework for modelling choice under uncertainty. In particular, they are able to represent the set-valued choices that appear in imprecise-probabilistic decision making. We provide these choice functions with a clear interpretation in terms of desirability, use this interpretation to derive a set of basic coherence axioms, and show that this notion of coherence leads to a representation in terms of sets of strict preference orders. By imposing additional properties such as totality, the mixing property and Archimedeanity, we obtain representation in terms of sets of strict total orders, lexicographic probability systems, coherent lower previsions or linear previsions.

Choice functions constitute a simple, direct and very general mathematical framework for modelling choice under uncertainty. In particular, they are able to represent the set-valued choices that typically arise from applying decision rules to imprecise-probabilistic uncertainty models. We provide them with a clear interpretation in terms of attitudes towards gambling, borrowing ideas from the theory of sets of desirable gambles, and we use this interpretation to derive a set of basic axioms. We show that these axioms lead to a full-fledged theory of coherent choice functions, which includes a representation in terms of sets of desirable gambles, and a conservative inference method.