We introduce the Granite Embedding models, a family of encoder-based embedding models designed for retrieval tasks, spanning dense-retrieval and sparse-retrieval architectures, with both English and Multilingual capabilities. This report provides the technical details of training these highly effective 12 layer embedding models, along with their efficient 6 layer distilled counterparts. Extensive evaluations show that the models, developed with techniques like retrieval oriented pretraining, contrastive finetuning, knowledge distillation, and model merging significantly outperform publicly available models of similar sizes on both internal IBM retrieval and search tasks, and have equivalent performance on widely-used information retrieval benchmarks, while being trained on high-quality data suitable for enterprise use. We publicly release all our Granite Embedding models under the Apache 2.0 license, allowing both research and commercial use at https://huggingface.co/collections/ibm-granite . Figure 1: Average performance on the Granite embedding models (in blue) vs BGE, GTE, Snowflake, E5, and Nomic models on 5 QA and IR datasets: BEIR, ClapNQ, CoIR, RedHat, and UnifiedSearch (the last 2 are internal IBM datasets). The goal of text embedding models is to convert variable length text into a fixed vector, encoding the text semantics into a multidimensional vector in such a way that semantically close texts are close in the vector space, while dissimilar texts have a low similarity. These embeddings can then be used in a variety of tasks, most commonly in retrieval applications, where the relevance of a document to a given query can be determined by the similarity of their embeddings (Dunn et al., 2017; Xiong et al., 2020; Neelakantan et al., 2022)(Zamani et al., 2018; Zhao et al., 2020), but also in document clustering (Angelov, 2020) and text classification (Sun et al., 2019). See Contributions section for full author list.

Current Large Language Model (LLM) agents demonstrate strong reasoning and tool use capabilities, but often lack self-awareness, failing to balance these approaches effectively. This imbalance leads to Tool Overuse, where models unnecessarily rely on external tools for tasks solvable with parametric knowledge, increasing computational overhead. Inspired by human metacognition, we introduce SMART (Strategic Model-Aware Reasoning with Tools), a paradigm that enhances an agent's self-awareness to optimize task handling and reduce tool overuse. To support this paradigm, we introduce SMART-ER, a dataset spanning three domains, where reasoning alternates between parametric knowledge and tool-dependent steps, with each step enriched by rationales explaining when tools are necessary. Through supervised training, we develop SMARTAgent, a family of models that dynamically balance parametric knowledge and tool use. Evaluations show that SMARTAgent reduces tool use by 24% while improving performance by over 37%, enabling 7B-scale models to match its 70B counterpart and GPT-4o. Additionally, SMARTAgent generalizes to out-of-distribution test data like GSM8K and MINTQA, maintaining accuracy with just one-fifth the tool calls. These highlight the potential of strategic tool use to enhance reasoning, mitigate overuse, and bridge the gap between model size and performance, advancing intelligent and resource-efficient agent designs.

This white paper, developed through close collaboration between IBM Research and UIUC researchers within the IIDAI Institute, envisions transforming hybrid cloud systems to meet the growing complexity of AI workloads through innovative, full-stack co-design approaches, emphasizing usability, manageability, affordability, adaptability, efficiency, and scalability. By integrating cutting-edge technologies such as generative and agentic AI, cross-layer automation and optimization, unified control plane, and composable and adaptive system architecture, the proposed framework addresses critical challenges in energy efficiency, performance, and cost-effectiveness. Incorporating quantum computing as it matures will enable quantum-accelerated simulations for materials science, climate modeling, and other high-impact domains. Collaborative efforts between academia and industry are central to this vision, driving advancements in foundation models for material design and climate solutions, scalable multimodal data processing, and enhanced physics-based AI emulators for applications like weather forecasting and carbon sequestration. Research priorities include advancing AI agentic systems, LLM as an Abstraction (LLMaaA), AI model optimization and unified abstractions across heterogeneous infrastructure, end-to-end edge-cloud transformation, efficient programming model, middleware and platform, secure infrastructure, application-adaptive cloud systems, and new quantum-classical collaborative workflows. These ideas and solutions encompass both theoretical and practical research questions, requiring coordinated input and support from the research community. This joint initiative aims to establish hybrid clouds as secure, efficient, and sustainable platforms, fostering breakthroughs in AI-driven applications and scientific discovery across academia, industry, and society.

We develop a method for training small-scale (under 100M parameter) neural information retrieval models with as few as 10 gold relevance labels. The method depends on generating synthetic queries for documents using a language model (LM), and the key step is that we automatically optimize the LM prompt that is used to generate these queries based on training quality. In experiments with the BIRCO benchmark, we find that models trained with our method outperform RankZephyr and are competitive Figure 1: An overview of the PATH pipeline for training with RankLLama, both of which are 7B parameter a reranker with synthetic queries. A user only needs to models trained on over 100K labels. These input a prompt with the task description and as few as findings point to the power of automatic prompt 10 relevance judgements to achieve strong results.

Large (NQ) (Kwiatkowski et al., 2019) and SQuAD (Rajpurkar scale research in this area began with the tasks et al., 2016, 2018) which are just a few of Machine Reading Comprehension (Rajpurkar words. It is grounded on a single gold passage, et al., 2016; Rogers et al., 2023; Fisch et al., in contrast to other long-form question answering 2021), and Information Retrieval (Manning et al., (LFQA) datasets such as ELI5 (Fan et al., 2019) 2008; Voorhees and Harman, 2005; Thakur et al., where gold passages are not available. It is built 2021) and has more recently been come to be from a subset of the highly successful Natural Questions known as Retrieval Augmented Generation (Lewis (Kwiatkowski et al., 2019) dataset for extractive et al., 2021; Guu et al., 2020) which encompasses QA from Wikipedia documents based on users both tasks. The recent popularity of generative real web search queries - specifically, the subset of AI with Large Language models (LLM), such as NQ that has long answers (passages) but no short GPT (Brown et al., 2020), Llama (Touvron et al., extractive answers.

We present a large-scale empirical study of how choices of configuration parameters affect performance in knowledge distillation (KD). An example of such a KD parameter is the measure of distance between the predictions of the teacher and the student, common choices for which include the mean squared error (MSE) and the KL-divergence. Although scattered efforts have been made to understand the differences between such options, the KD literature still lacks a systematic study on their general effect on student performance. We take an empirical approach to this question in this paper, seeking to find out the extent to which such choices influence student performance across 13 datasets from 4 NLP tasks and 3 student sizes. We quantify the cost of making sub-optimal choices and identify a single configuration that performs well across the board.

Offensive language such as hate, abuse, and profanity (HAP) occurs in various content on the web. While previous work has mostly dealt with sentence level annotations, there have been a few recent attempts to identify offensive spans as well. We build upon this work and introduce Muted, a system to identify multilingual HAP content by displaying offensive arguments and their targets using heat maps to indicate their intensity. Muted can leverage any transformer-based HAP-classification model and its attention mechanism out-of-the-box to identify toxic spans, without further fine-tuning. In addition, we use the spaCy library to identify the specific targets and arguments for the words predicted by the attention heatmaps. We present the model's performance on identifying offensive spans and their targets in existing datasets and present new annotations on German text. Finally, we demonstrate our proposed visualization tool on multilingual inputs.

Despite their remarkable capabilities, large language models (LLMs) often produce responses containing factual inaccuracies due to their sole reliance on the parametric knowledge they encapsulate. Retrieval-Augmented Generation (RAG), an ad hoc approach that augments LMs with retrieval of relevant knowledge, decreases such issues. However, indiscriminately retrieving and incorporating a fixed number of retrieved passages, regardless of whether retrieval is necessary, or passages are relevant, diminishes LM versatility or can lead to unhelpful response generation. We introduce a new framework called Self-Reflective Retrieval-Augmented Generation (Self-RAG) that enhances an LM's quality and factuality through retrieval and self-reflection. Our framework trains a single arbitrary LM that adaptively retrieves passages on-demand, and generates and reflects on retrieved passages and its own generations using special tokens, called reflection tokens. Generating reflection tokens makes the LM controllable during the inference phase, enabling it to tailor its behavior to diverse task requirements. Experiments show that Self-RAG (7B and 13B parameters) significantly outperforms state-of-the-art LLMs and retrieval-augmented models on a diverse set of tasks. Specifically, Self-RAG outperforms ChatGPT and retrieval-augmented Llama2-chat on Open-domain QA, reasoning and fact verification tasks, and it shows significant gains in improving factuality and citation accuracy for long-form generations relative to these models.

Many information retrieval tasks require large labeled datasets for fine-tuning. However, such datasets are often unavailable, and their utility for real-world applications can diminish quickly due to domain shifts. To address this challenge, we develop and motivate a method for using large language models (LLMs) to generate large numbers of synthetic queries cheaply. The method begins by generating a small number of synthetic queries using an expensive LLM. After that, a much less expensive one is used to create large numbers of synthetic queries, which are used to fine-tune a family of reranker models. These rerankers are then distilled into a single efficient retriever for use in the target domain. We show that this technique Figure 1: Overview of UDAPDR. An expensive LLM boosts zero-shot accuracy in long-tail domains like GPT-3 is used to create an initial set of synthetic and achieves substantially lower latency than queries. These are incorporated into a set of prompts for standard reranking methods.

Neural information retrieval often adopts a retrieve-and-rerank framework: a bi-encoder network first retrieves K (e.g., 100) candidates that are then re-ranked using a more powerful cross-encoder model to rank the better candidates higher. The re-ranker generally produces better candidate scores than the retriever, but is limited to seeing only the top K retrieved candidates, thus providing no improvements in retrieval performance as measured by Recall@K. In this work, we leverage the re-ranker to also improve retrieval by providing inference-time relevance feedback to the retriever. Concretely, we update the retriever's query representation for a test instance using a lightweight inference-time distillation of the re-ranker's prediction for that instance. The distillation loss is designed to bring the retriever's candidate scores closer to those of the re-ranker. A second retrieval step is then performed with the updated query vector. We empirically show that our approach, which can serve arbitrary retrieve-and-rerank pipelines, significantly improves retrieval recall in multiple domains, languages, and modalities.