Online learning to rank sequentially recommends a small list of items to users from a large candidate set and receives the users' click feedback. In many real-world scenarios, users browse the recommended list in order and click the first attractive item without checking the rest. Such behaviors are usually formulated as the cascade model. Many recent works study algorithms for cascading bandits, an online learning to rank framework in the cascade model. However, the performance of existing methods may drop significantly if part of the user feedback is adversarially corrupted (e.g., click fraud). In this work, we study how to resist adversarial corruptions in cascading bandits. We first formulate the ``\textit{Cascading Bandits with Adversarial Corruptions}" (CBAC) problem, which assumes that there is an adaptive adversary that may manipulate the user feedback. Then we propose two robust algorithms for this problem, which assume the corruption level is known and agnostic, respectively. We show that both algorithms can achieve logarithmic regret when the algorithm is not under attack, and the regret increases linearly with the corruption level. The experimental results also verify the robustness of our methods.
The combinatorial multi-armed bandit (CMAB) is a fundamental sequential decision-making framework, extensively studied over the past decade. However, existing work primarily focuses on the online setting, overlooking the substantial costs of online interactions and the readily available offline datasets. To overcome these limitations, we introduce Off-CMAB, the first offline learning framework for CMAB. Central to our framework is the combinatorial lower confidence bound (CLCB) algorithm, which combines pessimistic reward estimations with combinatorial solvers. To characterize the quality of offline datasets, we propose two novel data coverage conditions and prove that, under these conditions, CLCB achieves a near-optimal suboptimality gap, matching the theoretical lower bound up to a logarithmic factor. We validate Off-CMAB through practical applications, including learning to rank, large language model (LLM) caching, and social influence maximization, showing its ability to handle nonlinear reward functions, general feedback models, and out-of-distribution action samples that excludes optimal or even feasible actions. Extensive experiments on synthetic and real-world datasets further highlight the superior performance of CLCB.
In recent years, there has been a growing interest in using machine learning (ML) in query optimization to select more efficient plans. Existing learning-based query optimizers use certain model architectures to convert tree-structured query plans into representations suitable for downstream ML tasks. As the design of these architectures significantly impacts cost estimation, we propose a tree model architecture based on Bidirectional Graph Neural Networks (Bi-GNN) aggregated by Gated Recurrent Units (GRUs) to achieve more accurate cost estimates. The inherent uncertainty of data and model parameters also leads to inaccurate cost estimates, resulting in suboptimal plans and less robust query performance. To address this, we implement a novel learning-to-rank cost model that effectively quantifies the uncertainty in cost estimates using approximate probabilistic ML. This model adaptively integrates quantified uncertainty with estimated costs and learns from comparing pairwise plans, achieving more robust performance. In addition, we propose the first explainability technique specifically designed for learning-based cost models. This technique explains the contribution of any subgraphs in the query plan to the final predicted cost, which can be integrated and trained with any learning-based cost model to significantly boost the model's explainability. By incorporating these innovations, we propose a cost model for a Robust and Explainable Query Optimizer, Reqo, that improves the accuracy, robustness, and explainability of cost estimation, outperforming state-of-the-art approaches in all three dimensions.
Multi-modal retrieval-augmented Question Answering (MRAQA), integrating text and images, has gained significant attention in information retrieval (IR) and natural language processing (NLP). Traditional ranking methods rely on small encoder-based language models, which are incompatible with modern decoder-based generative large language models (LLMs) that have advanced various NLP tasks. To bridge this gap, we propose RAMQA, a unified framework combining learning-to-rank methods with generative permutation-enhanced ranking techniques. We first train a pointwise multi-modal ranker using LLaVA as the backbone. Then, we apply instruction tuning to train a LLaMA model for re-ranking the top-k documents using an innovative autoregressive multi-task learning approach. Our generative ranking model generates re-ranked document IDs and specific answers from document candidates in various permutations. Experiments on two MRAQA benchmarks, WebQA and MultiModalQA, show significant improvements over strong baselines, highlighting the effectiveness of our approach. Code and data are available at: https://github.com/TonyBY/RAMQA
A review of over 160,000 customer cases indicates that about 90% of time is spent by the product support for solving around 10% of subset of tickets where a trivial solution may not exist. Many of these challenging cases require the support of several engineers working together within a "swarm", and some also need to go to development support as bugs. These challenging customer issues represent a major opportunity for machine learning and knowledge graph that identifies the ideal engineer / group of engineers(swarm) that can best address the solution, reducing the wait times for the customer. The concrete ML task we consider here is a learning-to-rank(LTR) task that given an incident and a set of engineers currently assigned to the incident (which might be the empty set in the non-swarming context), produce a ranked list of engineers best fit to help resolve that incident. To calculate the rankings, we may consider a wide variety of input features including the incident description provided by the customer, the affected component(s), engineer ratings of their expertise, knowledge base article text written by engineers, response to customer text written by engineers, and historic swarming data. The central hypothesis test is that by including a holistic set of contextual data around which cases an engineer has solved, we can significantly improve the LTR algorithm over benchmark models. The article proposes a novel approach of modelling Knowledge Graph embeddings from multiple data sources, including the swarm information. The results obtained proves that by incorporating this additional context, we can improve the recommendations significantly over traditional machine learning methods like TF-IDF.
Vision language models (VLMs) like CLIP show stellar zero-shot capability on classification benchmarks. However, selecting the VLM with the highest performance on the unlabeled downstream task is non-trivial. Existing VLM selection methods focus on the class-name-only setting, relying on a supervised large-scale dataset and large language models, which may not be accessible or feasible during deployment. This paper introduces the problem of \textbf{unsupervised vision-language model selection}, where only unsupervised downstream datasets are available, with no additional information provided. To solve this problem, we propose a method termed Visual-tExtual Graph Alignment (VEGA), to select VLMs without any annotations by measuring the alignment of the VLM between the two modalities on the downstream task. VEGA is motivated by the pretraining paradigm of VLMs, which aligns features with the same semantics from the visual and textual modalities, thereby mapping both modalities into a shared representation space. Specifically, we first construct two graphs on the vision and textual features, respectively. VEGA is then defined as the overall similarity between the visual and textual graphs at both node and edge levels. Extensive experiments across three different benchmarks, covering a variety of application scenarios and downstream datasets, demonstrate that VEGA consistently provides reliable and accurate estimates of VLMs' performance on unlabeled downstream tasks.
Online Learning to Rank (OLTR) optimises ranking models using implicit user feedback, such as clicks. Unlike traditional Learning to Rank (LTR) methods that rely on a static set of training data with relevance judgements to learn a ranking model, OLTR methods update the model continually as new data arrives. Thus, it addresses several drawbacks such as the high cost of human annotations, potential misalignment between user preferences and human judgments, and the rapid changes in user query intents. However, OLTR methods typically require the collection of searchable data, user queries, and clicks, which poses privacy concerns for users. Federated Online Learning to Rank (FOLTR) integrates OLTR within a Federated Learning (FL) framework to enhance privacy by not sharing raw data. While promising, FOLTR methods currently lag behind traditional centralised OLTR due to challenges in ranking effectiveness, robustness with respect to data distribution across clients, susceptibility to attacks, and the ability to unlearn client interactions and data. This thesis presents a comprehensive study on Federated Online Learning to Rank, addressing its effectiveness, robustness, security, and unlearning capabilities, thereby expanding the landscape of FOLTR.
Developing increasingly efficient and accurate algorithms for approximate nearest neighbor search is a paramount goal in modern information retrieval. A primary approach to addressing this question is clustering, which involves partitioning the dataset into distinct groups, with each group characterized by a representative data point. By this method, retrieving the top-k data points for a query requires identifying the most relevant clusters based on their representatives -- a routing step -- and then conducting a nearest neighbor search within these clusters only, drastically reducing the search space. The objective of this thesis is not only to provide a comprehensive explanation of clustering-based approximate nearest neighbor search but also to introduce and delve into every aspect of our novel state-of-the-art method, which originated from a natural observation: The routing function solves a ranking problem, making the function amenable to learning-to-rank. The development of this intuition and applying it to maximum inner product search has led us to demonstrate that learning cluster representatives using a simple linear function significantly boosts the accuracy of clustering-based approximate nearest neighbor search.
Recent advances in vision-language models (VLMs) have made significant progress in downstream tasks that require quantitative concepts such as facial age estimation and image quality assessment, enabling VLMs to explore applications like image ranking and retrieval. However, existing studies typically focus on the reasoning based on a single image and heavily depend on text prompting, limiting their ability to learn comprehensive understanding from multiple images. To address this, we propose an effective yet efficient approach that reframes the CLIP model into a learning-to-rank task and introduces a lightweight adapter to augment CLIP for text-guided image ranking. Specifically, our approach incorporates learnable prompts to adapt to new instructions for ranking purposes and an auxiliary branch with ranking-aware attention, leveraging text-conditioned visual differences for additional supervision in image ranking. Our ranking-aware adapter consistently outperforms fine-tuned CLIPs on various tasks and achieves competitive results compared to state-of-the-art models designed for specific tasks like facial age estimation and image quality assessment. Overall, our approach primarily focuses on ranking images with a single instruction, which provides a natural and generalized way of learning from visual differences across images, bypassing the need for extensive text prompts tailored to individual tasks. Code is available: https://github.com/uynaes/RankingAwareCLIP.
We propose a novel approach to learn relational policies for classical planning based on learning to rank actions. We introduce a new graph representation that explicitly captures action information and propose a Graph Neural Network architecture augmented with Gated Recurrent Units (GRUs) to learn action rankings. Our model is trained on small problem instances and generalizes to significantly larger instances where traditional planning becomes computationally expensive. Experimental results across standard planning benchmarks demonstrate that our action-ranking approach achieves generalization to significantly larger problems than those used in training.
In e-commerce, ranking the search results based on users' preference is the most important task. Commercial e-commerce platforms, such as, Amazon, Alibaba, eBay, Walmart, etc. perform extensive and relentless research to perfect their search result ranking algorithms because the quality of ranking drives a user's decision to purchase or not to purchase an item, directly affecting the profitability of the e-commerce platform. In such a commercial platforms, for optimizing search result ranking numerous features are considered, which emerge from relevance, personalization, seller's reputation and paid promotion. To maintain their competitive advantage in the market, the platforms do no publish their core ranking algorithms, so it is difficult to know which of the algorithms or which of the features is the most effective for finding the most optimal search result ranking in e-commerce. No extensive surveys of ranking to rank in the e-commerce domain is also not yet published. In this work, we survey the existing e-commerce learning to rank algorithms. Besides, we also compare these algorithms based on query relevance criterion on a large real-life e-commerce dataset and provide a quantitative analysis. To the best of our knowledge this is the first such survey which include an experimental comparison among various learning to rank algorithms.
Influence function, a technique rooted in robust statistics, has been adapted in modern machine learning for a novel application: data attribution -- quantifying how individual training data points affect a model's predictions. However, the common derivation of influence functions in the data attribution literature is limited to loss functions that can be decomposed into a sum of individual data point losses, with the most prominent examples known as M-estimators. This restricts the application of influence functions to more complex learning objectives, which we refer to as non-decomposable losses, such as contrastive or ranking losses, where a unit loss term depends on multiple data points and cannot be decomposed further. In this work, we bridge this gap by revisiting the general formulation of influence function from robust statistics, which extends beyond M-estimators. Based on this formulation, we propose a novel method, the Versatile Influence Function (VIF), that can be straightforwardly applied to machine learning models trained with any non-decomposable loss. In comparison to the classical approach in statistics, the proposed VIF is designed to fully leverage the power of auto-differentiation, hereby eliminating the need for case-specific derivations of each loss function. We demonstrate the effectiveness of VIF across three examples: Cox regression for survival analysis, node embedding for network analysis, and listwise learning-to-rank for information retrieval. In all cases, the influence estimated by VIF closely resembles the results obtained by brute-force leave-one-out retraining, while being up to $10^3$ times faster to compute. We believe VIF represents a significant advancement in data attribution, enabling efficient influence-function-based attribution across a wide range of machine learning paradigms, with broad potential for practical use cases.
In-Context Learning (ICL) enables Large Language Models (LLMs) to perform new tasks by conditioning on prompts with relevant information. Retrieval-Augmented Generation (RAG) enhances ICL by incorporating retrieved documents into the LLM's context at query time. However, traditional retrieval methods focus on semantic relevance, treating retrieval as a search problem. In this paper, we propose reframing retrieval for ICL as a recommendation problem, aiming to select documents that maximize utility in ICL tasks. We introduce the In-Context Learning Embedding and Reranker Benchmark (ICLERB), a novel evaluation framework that compares retrievers based on their ability to enhance LLM accuracy in ICL settings. Additionally, we propose a novel Reinforcement Learning-to-Rank from AI Feedback (RLRAIF) algorithm, designed to fine-tune retrieval models using minimal feedback from the LLM. Our experimental results reveal notable differences between ICLERB and existing benchmarks, and demonstrate that small models fine-tuned with our RLRAIF algorithm outperform large state-of-the-art retrieval models. These findings highlight the limitations of existing evaluation methods and the need for specialized benchmarks and training strategies adapted to ICL.
Display advertising provides significant value to advertisers, publishers, and users. Traditional display advertising systems utilize a multi-stage architecture consisting of retrieval, coarse ranking, and final ranking. However, conventional retrieval methods rely on ID-based learning to rank mechanisms and fail to adequately utilize the content information of ads, which hampers their ability to provide diverse recommendation lists. To address this limitation, we propose leveraging the extensive world knowledge of LLMs. However, three key challenges arise when attempting to maximize the effectiveness of LLMs: "How to capture user interests", "How to bridge the knowledge gap between LLMs and advertising system", and "How to efficiently deploy LLMs". To overcome these challenges, we introduce a novel LLM-based framework called LLM Empowered Display ADvertisement REcommender system (LEADRE). LEADRE consists of three core modules: (1) The Intent-Aware Prompt Engineering introduces multi-faceted knowledge and designs intent-aware <Prompt, Response> pairs that fine-tune LLMs to generate ads tailored to users' personal interests. (2) The Advertising-Specific Knowledge Alignment incorporates auxiliary fine-tuning tasks and Direct Preference Optimization (DPO) to align LLMs with ad semantic and business value. (3) The Efficient System Deployment deploys LEADRE in an online environment by integrating both latency-tolerant and latency-sensitive service. Extensive offline experiments demonstrate the effectiveness of LEADRE and validate the contributions of individual modules. Online A/B test shows that LEADRE leads to a 1.57% and 1.17% GMV lift for serviced users on WeChat Channels and Moments separately. LEADRE has been deployed on both platforms, serving tens of billions of requests each day.
Traditional Learning-To-Rank (LETOR) approaches, including pairwise methods like RankNet and LambdaMART, often fall short by solely focusing on pairwise comparisons, leading to sub-optimal global rankings. Conversely, deep learning based listwise methods, while aiming to optimise entire lists, require complex tuning and yield only marginal improvements over robust pairwise models. To overcome these limitations, we introduce Travelling Salesman Problem Rank (TSPRank), a hybrid pairwise-listwise ranking method. TSPRank reframes the ranking problem as a Travelling Salesman Problem (TSP), a well-known combinatorial optimisation challenge that has been extensively studied for its numerous solution algorithms and applications. This approach enables the modelling of pairwise relationships and leverages combinatorial optimisation to determine the listwise ranking. This approach can be directly integrated as an additional component into embeddings generated by existing backbone models to enhance ranking performance. Our extensive experiments across three backbone models on diverse tasks, including stock ranking, information retrieval, and historical events ordering, demonstrate that TSPRank significantly outperforms both pure pairwise and listwise methods. Our qualitative analysis reveals that TSPRank's main advantage over existing methods is its ability to harness global information better while ranking. TSPRank's robustness and superior performance across different domains highlight its potential as a versatile and effective LETOR solution.
Recent advances have demonstrated that large language models (LLMs) excel as listwise rerankers, but their high computational demands remain a barrier to widespread adoption. Further, the traditional language modeling (LM) objective is not ideally suited for reranking tasks. FIRST is a novel approach that addresses these challenges by integrating a learning-to-rank objective and leveraging the logits of only the first generated token, thereby significantly reducing inference latency compared to traditional LLM rerankers. In this study, we extend the evaluation of FIRST to the TREC Deep Learning datasets (DL19-22), validating its robustness across diverse domains. We investigate the influence of different first-stage retrievers on FIRST rerankers, observing diminishing returns and patterns consistent with traditional LLM rerankers. Through applying the FIRST objective to a broader range of backbone models, we achieve effectiveness surpassing the original implementation. Our experiments confirm that fast reranking with single-token logits does not compromise out-of-domain reranking quality. To better quantify the computational savings in the original study, we measure and compare latency to find a 21%-42% gain across various models and benchmarks. Moreover, while LM training implicitly improves zero-shot single-token reranking, our experiments also raise questions about whether LM pre-training may hinder subsequent fine-tuning with the FIRST objective. These findings pave the way for more efficient and effective listwise reranking in future applications.
This study examines the potential of integrating Learning-to-Rank (LTR) with Query-focused Summarization (QFS) to enhance the summary relevance via content prioritization. Using a shared secondary decoder with the summarization decoder, we carry out the LTR task at the segment level. Compared to the state-of-the-art, our model outperforms on QMSum benchmark (all metrics) and matches on SQuALITY benchmark (2 metrics) as measured by Rouge and BertScore while offering a lower training overhead. Specifically, on the QMSum benchmark, our proposed system achieves improvements, particularly in Rouge-L (+0.42) and BertScore (+0.34), indicating enhanced understanding and relevance. While facing minor challenges in Rouge-1 and Rouge-2 scores on the SQuALITY benchmark, the model significantly excels in Rouge-L (+1.47), underscoring its capability to generate coherent summaries. Human evaluations emphasize the efficacy of our method in terms of relevance and faithfulness of the generated summaries, without sacrificing fluency. A deeper analysis reveals our model's superiority over the state-of-the-art for broad queries, as opposed to specific ones, from a qualitative standpoint. We further present an error analysis of our model, pinpointing challenges faced and suggesting potential directions for future research in this field.
State-of-the-art Extreme Multi-Label Text Classification (XMTC) models rely heavily on multi-label attention layers to focus on key tokens in input text, but obtaining optimal attention weights is challenging and resource-intensive. To address this, we introduce PLANT -- Pretrained and Leveraged AtteNTion -- a novel transfer learning strategy for fine-tuning XMTC decoders. PLANT surpasses existing state-of-the-art methods across all metrics on mimicfull, mimicfifty, mimicfour, eurlex, and wikiten datasets. It particularly excels in few-shot scenarios, outperforming previous models specifically designed for few-shot scenarios by over 50 percentage points in F1 scores on mimicrare and by over 36 percentage points on mimicfew, demonstrating its superior capability in handling rare codes. PLANT also shows remarkable data efficiency in few-shot scenarios, achieving precision comparable to traditional models with significantly less data. These results are achieved through key technical innovations: leveraging a pretrained Learning-to-Rank model as the planted attention layer, integrating mutual-information gain to enhance attention, introducing an inattention mechanism, and implementing a stateful-decoder to maintain context. Comprehensive ablation studies validate the importance of these contributions in realizing the performance gains.
Aligning large language models with human preferences improves interaction quality and safety by ensuring outputs better reflect human values. A promising strategy involves Reinforcement Learning from Human Feedback (RLHF), starting with collecting and ranking responses generated by a supervised fine-tuning model to refine alignment. Current methods (DPO) focus on learning from pairwise preference data, categorizing responses into preferred and less preferred pairs, and optimizing by maximizing pairwise margins. Recent studies have uncovered a substantial discrepancy between the theoretical aspirations of preference learning and its real-world results. Current preference alignment techniques underperform expectations, with ranking accuracies below $60\%$ on standard datasets. This suggests existing methods inadequately capture ideal preference relationships within sequences. To address this challenge, this paper introduces \underline{D}irect \underline{R}anking \underline{P}reference \underline{O}ptimization (DRPO), a novel method that views human preference alignment as a Learning-to-Rank (LTR) task. DRPO leverages NDCG, a widely used LTR metric, to optimize the ranking of responses within lists based on preference data, thereby enhancing ranking accuracies. Due to the nondifferentiability of NDCG, we propose diffNDCG loss, a differentiable approximation facilitated by a sorting network to simulate NDCG. Furthermore, to improve the quality of generated response, we propose a novel margin-based Adaptive Rank Policy Score. Extensive experiments have shown that DRPO outperforms existing baseline methods, enhancing the quality of the generated responses.
We introduce a novel framework called combinatorial logistic bandits (CLogB), where in each round, a subset of base arms (called the super arm) is selected, with the outcome of each base arm being binary and its expectation following a logistic parametric model. The feedback is governed by a general arm triggering process. Our study covers CLogB with reward functions satisfying two smoothness conditions, capturing application scenarios such as online content delivery, online learning to rank, and dynamic channel allocation. We first propose a simple yet efficient algorithm, CLogUCB, utilizing a variance-agnostic exploration bonus. Under the 1-norm triggering probability modulated (TPM) smoothness condition, CLogUCB achieves a regret bound of $\tilde{O}(d\sqrt{\kappa KT})$, where $\tilde{O}$ ignores logarithmic factors, $d$ is the dimension of the feature vector, $\kappa$ represents the nonlinearity of the logistic model, and $K$ is the maximum number of base arms a super arm can trigger. This result improves on prior work by a factor of $\tilde{O}(\sqrt{\kappa})$. We then enhance CLogUCB with a variance-adaptive version, VA-CLogUCB, which attains a regret bound of $\tilde{O}(d\sqrt{KT})$ under the same 1-norm TPM condition, improving another $\tilde{O}(\sqrt{\kappa})$ factor. VA-CLogUCB shows even greater promise under the stronger triggering probability and variance modulated (TPVM) condition, achieving a leading $\tilde{O}(d\sqrt{T})$ regret, thus removing the additional dependency on the action-size $K$. Furthermore, we enhance the computational efficiency of VA-CLogUCB by eliminating the nonconvex optimization process when the context feature map is time-invariant while maintaining the tight $\tilde{O}(d\sqrt{T})$ regret. Finally, experiments on synthetic and real-world datasets demonstrate the superior performance of our algorithms compared to benchmark algorithms.