Natural language processing (NLP) has witnessed impressive developments in answering questions, summarizing or translating reports, and analyzing sentiment or offensiveness. Much of this progress is owed to training ever-larger language models, such as T5 or GPT-3, that use deep monolithic architectures to internalize how language is used within text from massive Web crawls. During training, these models distill the facts they read into implicit knowledge, storing in their parameters not only the capacity to “understand” language tasks, but also highly abstract knowledge representations of entities, events, and facts the model needs for solving tasks.
Despite the well-publicized success of large language models, their black-box nature hinders key goals of NLP. In particular, existing large language models are generally:
Inefficient. Researchers continue to enlarge these models, leading to striking inefficiencies as the field already pushes past 1 trillion parameters. This imposes a considerable environmental impact and its costs exclude all but a few large organizations from the ability to train—or in many cases even deploy—such models.
Opaque. They encode “knowledge” into model weights, synthesizing what they manage to memorize from training examples. This makes it difficult to discern what sources—if any—the model uses to make a prediction, a concerning problem in practice as these models frequently generate fluent yet untrue statements.
Static. They are expensive to update. We cannot efficiently adapt a GPT model trained on, say, Wikipedia text from 2019 so it reflects the knowledge encoded in the 2021 Wikipedia—or the latest snapshot of the medical preprint server medRXiv. In practice, adaptation often necessitates expensive retraining or fine-tuning on the new corpus.
This post explores an emerging alternative, Retrieval-based NLP, in which models directly “search” for information in a text corpus to exhibit knowledge, leveraging the representational strengths of language models while addressing the challenges above. Such models—including REALM, RAG, ColBERT-QA, and Baleen—are already advancing the state of the art for tasks like answering open-domain questions and verifying complex claims, all with architectures that back their predictions with checkable sources while being 100–1000× smaller, and thus far cheaper to execute, than GPT-3. At Stanford, we have shown that improving the expressivity and supervision of scalable neural retrievers can lead to much stronger NLP systems: for instance, ColBERT-QA improves answer correctness on open-QA benchmarks by up to 16 EM points and Baleen improves the ability to check complex claims on HoVer, correctly and with provenance, by up to 42 percentage points against existing work.
As Figure 1 illustrates, retrieval-based NLP methods view tasks as “open-book” exams: knowledge is encoded explicitly in the form of a text corpus like Wikipedia, the medical literature, or a software’s API documentation. When solving a language task, the model learns to search for pertinent passages and to then use the retrieved information for crafting knowledgeable responses. In doing so, retrieval helps decouple the capacity that language models have for understanding text from how they store knowledge, leading to three key advantages.
Tackling Inefficiency. Retrieval-based models can be much smaller and faster, and thus more environmentally friendly. Unlike black-box language models, the parameters no longer need to store an ever-growing list of facts, as such facts can be retrieved. Instead, we can dedicate those parameters for processing language and solving tasks, leaving us with smaller models that are highly effective. For instance, ColBERT-QA achieves 47.8% EM on the open-domain Natural Questions task, whereas a fine-tuned T5-11B model (with 24x more parameters) and a few-shot GPT-3 model (with 400x more parameters) achieve only 34.8% and 29.9%, respectively.
Tackling Opaqueness. Retrieval-based NLP offers a transparent contract with users: when the model produces an answer, we can read the sources it retrieved and judge their relevance and credibility for ourselves. This is essential whether the model is factually correct or not: by inspecting the sources surfaced by a system like Baleen, we can trust its outputs only if we find that reliable sources do support them.
Tackling Static Knowledge. Retrieval-based models emphasize learning general techniques for finding and connecting information from the available resources. With facts stored as text, the retrieval knowledge store can be efficiently updated or expanded by modifying the text corpus, all while the model’s capacity for finding and using information remains constant. Besides computational cost reductions, this expedites generality: developers, even in niche domains, can “plug in” a domain-specific text collection and rely on retrieval to facilitate domain-aware responses.
ColBERT: Scalable yet expressive neural retrieval
As the name suggests, retrieval-based NLP relies on semantically rich search to extract information. For search be practical and effective, it must scale to massive text corpora. To draw on the open-book exam analogy, it’s hopeless to linearly look through the pages of a hefty textbook during the exam—we need scalable strategies for organizing the content in advance, and efficient techniques for locating relevant information at inference time.
Traditionally in IR, search tasks were conducted using bag-of-words models like BM25, which seek documents that contain the same tokens as the query. In 2019, search was revolutionized with BERT for ranking and its deployment in Google and Bing for Web search. The standard approach is illustrated in Figure 2(a). Each document is concatenated with the query, and both are fed jointly into a BERT model, fine-tuned to estimate relevance. BERT doubled the MRR@10 quality metric over BM25 on the popular MS MARCO Passage Ranking leaderboard, but it simultaneously posed a fundamental limitation: scoring each query–document pair requires billions of computational operations (FLOPs). As a result, BERT can only be used to re-rank the top-k (e.g., top-1000) documents already extracted by simpler methods like BM25, having no capacity to recover useful documents that bag-of-word search misses.
The key limitation of this approach is that it encodes queries and documents jointly. Many representation-similarity systems have been proposed to tackle this, some of which re-purpose BERT within the paradigm depicted in Figure 2(b). In these systems (like SBERT and ORQA, and more recently DPR and ANCE, every document in the corpus is fed into a BERT encoder that produces a dense vector meant to capture the semantics of the document. At search time, the query is encoded, separately, through another BERT encoder, and the top-k related documents are found using a dot product between the query and document vectors. By removing the expensive interactions between the query and the document, these models are able to scale far more efficiently than the approach in Figure 2(a).
Nonetheless, representation-similarity models suffer from an architectural bottleneck: they encode the query and document into coarse-grained representations and model relevance as a single dot product. This greatly diminishes quality compared with expensive re-rankers that model token-level interactions between the contents of queries and documents. Can we efficiently scale fine-grained, contextual interactions to a massive corpus, without compromising speed or quality? It turns out that the answer is “yes”, using a paradigm called late interaction, first devised in our ColBERT1 [code] model, which appeared at SIGIR 2020.
As depicted in Figure 2(c), ColBERT independently encodes queries and documents into fine-grained multi-vector representations. It then attempts to softly and contextually locate each query token inside the document: for each query embedding, it finds the most similar embedding in the document with a “MaxSim” operator and then sums up all of the MaxSims to score the document. “MaxSim” is a careful choice that allows us to index the document embeddings for Approximate Nearest Neighbor (ANN) search, enabling us to scale this rich interaction to millions of passages with latency on the order of tens of milliseconds. For instance, ColBERT can search over all passages in English Wikipedia in approximately 70 milliseconds per query. On MS MARCO Passage Ranking, ColBERT preserved the MRR@10 quality of BERT re-rankers while boosting recall@1k to nearly 97% against the official BM25 ranking’s recall@1k of just 81%.
Making neural retrievers more lightweight remains an active area of development, with models like DeepImpact that trade away some quality for extreme forms of efficiency and developments like BPR and quantized ColBERT that reduce the storage footprint by an order of magnitude while preserving the quality of DPR and ColBERT, respectively.
ColBERT-QA and Baleen: Specializing neural retrieval to complex tasks, with tracked provenance
While scaling expressive search mechanisms is critical, NLP models need more than just finding the right documents. In particular, we want NLP models to use retrieval to answer questions, fact-check claims, respond informatively in a conversation, or identify the sentiment of a piece of text. Many tasks of this kind—dubbed knowledge-intensive language tasks—are collected in the KILT benchmark. The most popular task is open-domain question answering (or Open-QA). Systems are given a question from any domain and must produce an answer, often by reference to the passages in a large corpus, as depicted in Figure 1(b).
Two popular models in this space are REALM and RAG, which rely on the ORQA and DPR retrievers discussed earlier. REALM and RAG jointly tune a retriever as well as a reader, a modeling component that consumes the retrieved documents and produces answers or responses. Take RAG as an example: its reader is a generative BART model, which attends to the passages while generating the target outputs. While they constitute important steps toward retrieval-based NLP, REALM and RAG suffer from two major limitations. First, they use the restrictive paradigm of Figure 2(b) for retrieval, thereby sacrificing recall: they are often unable to find relevant passages for conducting their tasks. Second, when training the retriever, REALM and RAG collect documents by searching for them inside the training loop and, to make this practical, they freeze the document encoder when fine-tuning, restricting the model’s adaptation to the task.
ColBERT-QA2 is an Open-QA system (published at TACL’21) that we built on top of ColBERT to tackle both problems. By adapting ColBERT’s expressive search to the task, ColBERT-QA finds useful passages for a larger fraction of the questions and thus enables the reader component to answer more questions correctly and with provenance. In addition, ColBERT-QA introduces relevance-guided supervision (RGS), a training strategy whose goal is to adapt a retriever like ColBERT to the specifics of an NLP task like Open-QA. RGS proceeds in discrete rounds, using the retriever trained in the previous round to collect “positive” passages that are likely useful for the reader—specifically, passages ranked highly by the latest version of the retriever and that also overlap with the gold answer of the question—and challenging “negative” passages. By converging to a high coverage of positive passages and by effectively sampling hard negatives, ColBERT-QA improves retrieval Success@20 by more than 5-, 5-, and 12-point gains on the open-domain QA settings of NaturalQuestions, TriviaQA, and SQuAD, and thus greatly improves downstream answer match.
A more sophisticated version of the Open-QA task is multi-hop reasoning, where systems must answer questions or verify claims by gathering information from multiple sources. Systems in this space, like GoldEn, MDR, and IRRR, find relevant documents and “hop” between them—often by running additional searches—to find all pertinent sources. While these models have demonstrated strong performance for two-hop tasks, scaling robustly to more hops is challenging as the search space grows exponentially.
To tackle this, our Baleen3 system (accepted as a Spotlight paper at NeurIPS’21) introduces a richer pipeline for multi-hop retrieval: after each retrieval “hop”, Baleen summarizes the pertinent information from the passages into a short context that is used to inform future hops. In doing so, Baleen controls the search space architecturally—obviating the need to explore each potential passage at every hop—without sacrificing recall. Baleen also extends ColBERT’s late interaction: it allows the representations of different documents to “focus” on distinct parts of the same query, as each of those documents in the corpus might satisfy a distinct aspect of the same complex query. As a result of its more deliberate architecture and its stronger retrieval modeling, Baleen saturates retrieval on the popular two-hop HotPotQA benchmark (raising answer-recall@20 from 89% by MDR to 96%) and dramatically improves performance on the harder four-hop claim verification benchmark HoVer, finding all required passages in 92% of the examples—up from just 45% for the official baseline and 75% for a many-hop flavor of ColBERT.
In these tasks, when our retrieval-based models make predictions, we can inspect their underlying sources and decide whether we can trust the answer. And when model errors stem from specific sources, those can be removed or edited, and making sure models are faithful to such edits is an active area of work.
Generalizing models to new domains with robust neural retrieval
In addition to helping with efficiency and transparency, retrieval approaches promise to make domain generalization and knowledge updates much easier in NLP. Exhibiting up-to-date, domain-specific knowledge is essential for many applications: you might want to answer questions over recent publications on COVID-19 or to develop a chatbot that guides customers to suitable products among those currently available in a fast-evolving inventory. For such applications, NLP models should be able to leverage any corpus provided to them, without having to train a new version of the model for each emerging scenario or domain.
While large language models are trained using plenty of data from the Web, this snapshot is:
Static. The Web evolves as the world does: Wikipedia articles reflect new elected officials, news articles describe current events, and scientific papers communicate new research. Despite this, a language model trained in 2020 has no way to learn about 2021 events, short of training and releasing a new version of the model.
Incomplete. Many topics are under-represented in Web crawls like C4 and The Pile. Suppose we seek to answer questions over the ACL papers published 2010–2021; there is no guarantee that The Pile contains all papers from the ACL Anthology a priori and there is no way to plug that in ad-hoc without additional training. Even when some ACL papers are present (e.g., through arXiv, which is included in The Pile), they form only a tiny sliver of the data, and it is difficult to reliably restrict the model to specifically those papers for answering NLP questions.
Public-only. Many applications hinge on private text, like internal company policies, in-house software documentation, copyrighted textbooks and novels, or personal email. Because models like GPT-3 never see such data in their training, they are fundamentally incapable of exhibiting knowledge pertaining to those topics without special re-training or fine-tuning.
With retrieval-based NLP, models learn effective ways to encode and extract information, allowing them to generalize to updated text, specialized domains, or private data without resorting to additional training. This suggests a vision where developers “plug in” their text corpus, like in-house software documentation, which is indexed by a powerful retrieval-based NLP model that can then answer questions, solve classification tasks, or generate summaries using the knowledge from the corpus, while always supporting its predictions with provenance from the corpus.
An exciting benchmark connected to this space is BEIR, which evaluates retrievers on their capacity for search “out-of-the-box” on unseen IR tasks, like Argument Retrieval, and in new domains, like the COVID-19 research literature. While retrieval offers a concrete mechanism for generalizing NLP models to new domains, not every IR model generalizes equally: the BEIR evaluations highlight the impact of modeling and supervision choices on generalization. For instance, due to its late interaction modeling, a vanilla off-the-shelf ColBERT retriever achieved the strongest recall of all competing IR models in the initial BEIR evaluations, outperforming the other off-the-shelf dense retrievers—namely, DPR, ANCE, SBERT, and USE-QA—on 13 out of 17 datasets. The BEIR benchmark continues to develop quickly, a recent addition being the TAS-B model, which advances a sophisticated supervision approach to distill ColBERT and BERT models into single-vector representations, inheriting much of their robustness in doing so. While retrieval allows rapid deployment in new domains, explicitly adapting retrieval to new scenarios is also possible. This is an active area of research, with work like QGen and AugDPR that generate synthetic questions and use those to explicitly fine-tune retrievers for targeting a new corpus.
Summary: Is retrieval “all you need”?
The black-box nature of large language models like T5 and GPT-3 makes them inefficient to train and deploy, opaque in their knowledge representations and in backing their claims with provenance, and static in facing a constantly evolving world and diverse downstream contexts. This post explores retrieval-based NLP, where models retrieve information pertinent to solving their tasks from a plugged-in text corpus. This paradigm allows NLP models to leverage the representational strengths of language models, while needing much smaller architectures, offering transparent provenance for claims, and enabling efficient updates and adaptation.
We surveyed much of the existing and emerging work in this space and highlighted some of our work at Stanford, including ColBERT for scaling up expressive retrieval to massive corpora via late interaction, ColBERT-QA for accurately answering open-domain questions by adapting high-recall retrieval to the task, and Baleen for solving tasks that demand information from several independent sources using a condensed retrieval architecture. We continue to actively maintain our code as open source.
Acknowledgments. We would like to thank Megha Srivastava and Drew A. Hudson for helpful comments and feedback on this blog post. We also thank Ashwin Paranjape, Xiang Lisa Li, and Sidd Karamcheti for valuable and insightful discussions.
Omar Khattab and Matei Zaharia. “ColBERT: Efficient and Effective Passage Search via Contextualized Late Interaction over BERT.” Proceedings of the 43rd International ACM SIGIR conference on research and development in Information Retrieval. 2020. ↩
Omar Khattab, Christopher Potts, Matei Zaharia; “Relevance-guided Supervision for OpenQA with ColBERT.” Transactions of the Association for Computational Linguistics 2021; 9 929–944. doi: https://doi.org/10.1162/tacl_a_00405 ↩
Omar Khattab, Christopher Potts, and Matei Zaharia. “Baleen: Robust Multi-Hop Reasoning at Scale via Condensed Retrieval.” (To appear at NeurIPS 2021.) arXiv preprint arXiv:2101.00436 (2021). ↩