A deep reinforcement learning approach for slate re-ranking in e-commerce.
problem Improving user satisfaction in e-commerce by optimizing the ranking of items in a slate.
method Generator and Critic approach, using reinforcement learning and a Full Slate Critic model.
result The Generator and Critic approach significantly outperforms existing methods in slate evaluation and efficiency.
Ranking is a central task in machine learning and information retrieval. In this task, it is especially important to present the user with a slate of items that is appealing as a whole. This in turn requires taking into account interactions between items, since intuitively, placing an item on the slate affects the deci…
New algorithms reduce slate bandit regret for large slates, outperforming existing methods.
problem Non-separable reward functions in slate bandits with many slates.
method Design of algorithms with sub-linear regret.
result Sub-linear regret with respect to the time horizon for large number of slates.
Generative model learns to create coherent slates from prompts.
problem Generating coherent slates from natural language prompts.
method DMSG, a generative framework based on diffusion models.
result DMSG outperforms baselines in relevance and diversity.
Scalable model for slate recommendation learns reward probabilities.
problem Scalable personalized slate recommendation in large action spaces.
method Probabilistic Rank and Reward (PRR) model combining reward, interaction, and rank.
result PRR outperforms existing methods and is scalable to large action spaces.
A new estimator reduces variance in slate bandit OPE.
problem Large action spaces in slate bandits cause high variance in OPE.
method Develops Latent IPS (LIPS) to optimize slate abstractions for low variance and bias.
result LIPS substantially outperforms existing estimators in scenarios with non-linear rewards and large slate spaces.
The conventional solution to the recommendation problem greedily ranks individual document candidates by prediction scores. However, this method fails to optimize the slate as a whole, and hence, often struggles to capture biases caused by the page layout and document interdepedencies. The slate recommendation problem …
A new recommender system uses slates and Thompson Sampling to improve diversity and click rates.
problem Recommending relevant content to users in a dynamic environment.
method Variational Bayesian Recurrent Neural Net, Thompson Sampling, hierarchical prior, in-slate Thompson Sampling.
result Improved click rates and diversity in recommended slates.
We learn hierarchical slate representations for collaborative filtering.
problem Building models for recommendation systems with hierarchical slates.
method Learning low-dimensional embeddings of hierarchical slates using recursive composition rules.
result Improved recommendation system performance on a real-world dataset.
Efficient algorithms for contextual slate bandits with limited adaptivity.
problem Contextual slate bandit problem with limited adaptivity.
method Proposed B-SlateGLinCB and RS-SlateGLinCB algorithms for batched and rarely-switching settings.
result Achieved regret bounds of O(Nd^(3/2)√T) and O(Nd√T) under diversity assumption.
New estimator reduces risk in slate bandits by leveraging Bayes risk criterion.
problem Evaluating slate policies using logged data when policies factorize over slots.
method Developed a new estimator using a control variate approach, showing risk improvement over existing methods.
result The new estimator has lower risk than the pseudoinverse estimator in slate bandit problems.
Most practical recommender systems focus on estimating immediate user engagement without considering the long-term effects of recommendations on user behavior. Reinforcement learning (RL) methods offer the potential to optimize recommendations for long-term user engagement. However, since users are often presented with…
New algorithms learn MNL weights efficiently for any slate size.
problem Efficiently learn weights for MNL models given query access.
method Two algorithms: adaptive and non-adaptive, with specific query complexities.
result Optimal query complexities for both adaptive and non-adaptive cases.
We implement a method for re-ranking top-10 results of a state-of-the-art question answering (QA) system. The goal of our re-ranking approach is to improve the answer selection given the user question and the top-10 candidates. We focus on improving deployed QA systems that do not allow re-training or re-training comes…
This paper optimizes slate decision systems for large action spaces.
problem Optimizing large-scale decision systems with arbitrary reward functions.
method A policy optimization framework with a novel relaxation of decision functions.
result Demonstrates the effectiveness of the proposed method on large action spaces.
This paper improves fairness in recommendation systems by learning individual preferences across multiple dimensions.
problem Fairness in recommender systems, especially in areas with social impact.
method Opportunistic multi-aspect re-ranking approach that learns individual preferences and enhances provider fairness.
result Achieves a better trade-off between accuracy and fairness across multiple fairness dimensions.
A new dataset tracks user interactions and click responses in online marketplaces.
problem Lack of exposure data in recommender systems datasets.
method Proposes a novel dataset including slates and click responses, allowing more accurate likelihood models.
result Models using exposure data show more natural likelihood, reducing bias towards previously exposed items.
We analyze different re-ranking algorithms for diversification and show that majority of them are based on maximizing submodular/modular functions from the class of parameterized concave/linear over modular functions. We study the optimality of such algorithms in terms of the `total curvature'. We also show that by adj…
LITE models improve query-document relevance with learnable late interactions.
problem Improving query-document relevance with lower latency and storage.
method Proposes learnable late-interaction models (LITE) that use factorized query and document embeddings followed by a learnable scorer.
result Empirically, LITE outperforms previous late-interaction models in re-ranking tasks.
EBR improves NMT by re-ranking samples drawn from MLE-trained models.
problem Discrepancy between MLE and BLEU score in neural machine translation.
method Train an energy-based model to mimic BLEU score, then use it for re-ranking.
result EBR consistently improves NMT performance by +4 BLEU points on IWSLT'14 German-English.
Enhances currency strategy Sharpe ratio by 30% using context-aware Learning to Rank.
problem Sub-optimal ranking of assets during critical market periods.
method Context-aware Learning to Rank model based on Transformer architecture.
result Significantly improves Sharpe ratio and various performance metrics.
Meta-AAD uses deep reinforcement learning to improve anomaly detection by selecting the most informative instances.
problem High false-positive rate in anomaly detection, especially in high-stake applications.
method Meta-AAD leverages deep reinforcement learning to train a meta-policy for query selection, optimizing the number of discovered anomalies.
result Meta-AAD significantly outperforms state-of-the-art re-ranking strategies and unsupervised baselines on 24 benchmark datasets.
This paper studies the evaluation of policies that recommend an ordered set of items (e.g., a ranking) based on some context---a common scenario in web search, ads, and recommendation. We build on techniques from combinatorial bandits to introduce a new practical estimator that uses logged data to estimate a policy's p…
In this paper, we study the problem of safe online learning to re-rank, where user feedback is used to improve the quality of displayed lists. Learning to rank has traditionally been studied in two settings. In the offline setting, rankers are typically learned from relevance labels created by judges. This approach has…
New method for evaluating sequential recommendations with lower variance.
problem Evaluating good sequences of music, video, news, and e-commerce recommendations.
method Proposes a new counterfactual estimator for sequential reward interactions with lower variance and asymptotic unbiasedness.
result Our method outperforms existing methods in bias and data efficiency for sequential track recommendations.
Parallel deep learning architectures like fine-tuned BERT and MT-DNN, have quickly become the state of the art, bypassing previous deep and shallow learning methods by a large margin. More recently, pre-trained models from large related datasets have been able to perform well on many downstream tasks by just fine-tunin…
A neural architecture learns and refines graph correspondences.
problem Learning structural correspondences between graphs.
method Two-stage neural architecture: node embeddings followed by message passing networks.
result Message passing scheme computes well-founded consensus for corresponding neighborhoods.
The surfaces of many cultural heritage objects were embellished with various patterns, especially curve patterns. In practice, most of the unearthed cultural heritage objects are highly fragmented, e.g., sherds of potteries or vessels, and each of them only shows a very small portion of the underlying full design, with…
The method of random projections has become a standard tool for machine learning, data mining, and search with massive data at Web scale. The effective use of random projections requires efficient coding schemes for quantizing (real-valued) projected data into integers. In this paper, we focus on a simple 2-bit coding …
We present a deep neural-network model for lifelong learning inspired by several forms of neuroplasticity. The neural network develops continuously in response to signals from the environment. In the beginning, the network is a blank slate with no nodes at all. It develops according to four rules: (i) expansion, which …
Image captioning models are usually evaluated on their ability to describe a held-out set of images, not on their ability to generalize to unseen concepts. We study the problem of compositional generalization, which measures how well a model composes unseen combinations of concepts when describing images. State-of-the-…
We summarize the potential impact that the European Union's new General Data Protection Regulation will have on the routine use of machine learning algorithms. Slated to take effect as law across the EU in 2018, it will restrict automated individual decision-making (that is, algorithms that make decisions based on user…
Recommendations are broadly used in marketplaces to match users with items relevant to their interests and needs. To understand user intent and tailor recommendations to their needs, we use deep learning to explore various heterogeneous data available in marketplaces. This paper focuses on the challenge of measuring re…
Study efficient graph optimization with noisy data.
problem Optimizing functions on graphs with noisy observations.
method Best-arm identification and simulated annealing variants.
result Near-optimal solutions found with small query numbers.
Neural nets predict user attention from mouse movements.
problem Predicting user attention from mouse cursor movements.
method Investigated different mouse movement representations and trained neural networks.
result Neural networks outperform handcrafted features for predicting user attention.
Researchers show mixtures of ranking models are generally identifiable.
problem Understanding when and how parameters of mixtures of ranking models can be uniquely determined.
method Algebraic geometry framework applied to verify the number of solutions in polynomial systems.
result Popular mixtures of ranking models with two components are generically identifiable.
Search queries are appropriate when users have explicit intent, but they perform poorly when the intent is difficult to express or if the user is simply looking to be inspired. Visual browsing systems allow e-commerce platforms to address these scenarios while offering the user an engaging shopping experience. Here we …
This paper improves retrieval for LLMs in financial document Q&A.
problem Suboptimal text chunk retrieval by RAG causes inaccuracies in LLM responses.
method Sophisticated chunking techniques, query expansion, metadata annotations, re-ranking algorithms, and embedding fine-tuning.
result Enhanced retrieval quality improves LLM performance and reliability.
PG-EVIKAL refines molecular property predictions using neighbor fusion and evidential neural networks.
problem Improving molecular property predictions using test-time neighbor fusion.
method Adapting evidential neural networks to refine predictions by re-ranking structurally similar neighbors.
result PG-EVIKAL reduces RMSE on 14 out of 16 molecular datasets, improving calibration and sequential refinement.
PoPCoin aims to create a more equitable cryptocurrency.
problem Inequality in traditional money systems.
method Develops two rules for PoPCoin: equal distribution and demurrage.
result PoPCoin can limit monetary inequality and incentivize rapid growth.
Systems rank PubMed abstracts and sentences for RDoC criteria, achieving high mAP and MAA.
problem Lack of RDoC labeled datasets and complex labelling process hinder full use of RDoC framework.
method Attention-based neural topic models, supervised and unsupervised sentence ranking models, BM25, BoW, TF-IDF.
result Best systems achieved 1st rank with 0.86 mAP and 0.58 MAA.
Unified transformer-based LT-TTD improves ranking efficiency and quality.
problem Decoupled L1 and L2 models in recommendation and search systems cause irreversible error propagation and suboptimal ranking.
method LT-TTD combines two-tower models with transformer expressivity in a unified listwise learning framework, providing theoretical guarantees and UPQE evaluation.
result LT-TTD reduces irretrievable relevant items and achieves better global optimization than disjoint training.
Hybrid method uses LLM to filter lead-lag relationships in prediction markets.
problem Challenges in discovering robust lead-lag relationships in prediction markets due to spurious correlations.
method Two-stage approach: statistical Granger causality followed by LLM semantic re-ranking.
result LLM-based method outperforms statistical baseline, increasing win rate and reducing average loss magnitude.
Transformer models improve query-document retrieval efficiency and accuracy.
problem Efficiently retrieve relevant documents from large corpora for query matching.
method Designed paragraph-level pre-training tasks to optimize embedding-based Transformer models.
result Transformer models significantly outperform BM-25 and non-Transformer embedding models.
A new method reduces variance in training early-stage rankers for large-scale search systems.
problem Training early-stage rankers for large-scale search systems is challenging due to exploding variance in policy gradient methods.
method Proposes credit-assigned policy gradient (CA-PG) to mitigate variance in training early-stage rankers.
result CA-PG significantly reduces variance in training early-stage rankers compared to vanilla policy gradient.
The study examines how permutation-based optimization performance varies across different function representations.
problem Understanding how the order of function evaluations affects optimization performance.
method Iterative search setting with sampling without replacement, algebraic function recombination, correlation analysis, hierarchical clustering, PCA, ANOVA.
result Algebraically modified benchmarks yield stable re-rankings and coherent clusters of functions and sampling policies, indicating non-additive search effort.
In this paper we study the problems of estimating heterogeneity in causal effects in experimental or observational studies and conducting inference about the magnitude of the differences in treatment effects across subsets of the population. In applications, our method provides a data-driven approach to determine which…