The Case Against Generation for Retrieval: Discriminative Language Models as Effective Retrievers

Large Language Models (LLMs) have emerged as powerful assets for recommender systems. However, deploying them as generative recommenders or zero-shot rankers at web-scale remains bottlenecked by prohibitive computational overhead and grounding challenges. In this paper, we revitalize the classic, highly efficient two-tower retrieval architecture by adapting LLMs as semantic representation backbones rather than generative engines. We introduce an LLM-native two-tower framework engineered for high-throughput, large-scale retrieval. Our architecture introduces several key innovations: a shared LLM encoder for joint user-item modeling, End-Of-Sentence (EOS) token pooling for compact sequence embedding, cross-dataset transfer learning, knowledge distillation from powerful cross-encoder teachers, and latent reasoning within the user tower. Extensive evaluation across three public benchmarks demonstrates that cross-encoder architecture outperforms current state-of-the-art (SoTA) models, while the efficient two-tower student achieves SoTA-comparable retrieval performance. Furthermore, experiments on internal large-scale production systems yield substantial topline retrieval improvements along with high resilience to model staleness and superior data scaling. Our findings demonstrate that when augmented with modern representation learning, the traditional two-tower paradigm remains an exceptionally competitive and practical solution for industrial retrieval systems.

GR2 Technical Report

Industrial recommendation systems serve billions of users through a multi-stage funnel -- retrieval, early-stage ranking, and re-ranking -- where the final re-ranking step disproportionately shapes user engagement and downstream performance, particularly for carousel and grid display formats. Despite growing enthusiasm for Large Language Models (LLMs) in recommendation, three gaps hinder industrial adoption: (1) most efforts target retrieval and ranking, leaving re-ranking -- the stage closest to the final user experience -- largely underexplored; (2) LLMs are typically deployed zero-shot or via supervised fine-tuning, underutilizing the reasoning capabilities unlocked by reinforcement learning (RL) on verifiable rewards; (3) deployed catalogs index billions of items with non-semantic identifiers that lie outside any base-LLM vocabulary. We present GR2 (Generative Reasoning Re-Ranker), an end-to-end framework that combines (i) mid-training on semantic IDs produced by a tokenizer with >=99% uniqueness, (ii) reasoning-trace distilled from a stronger teacher via targeted prompting and rejection sampling, and (iii) RL with verifiable rewards purpose-built for re-ranking. To make GR2 resource-viable, we further (iv) introduce a context compressor that amortizes training cost, On-Policy Distillation (OPD) as a scalable alternative to SFT -- which we find collapses at industrial scale -- and reasoning distillation for low-latency serving. GR2 delivers +18.7% R@1, +7.1% R@3, and +9.6% N@3 over legacy baselines on industrial-scale traffic. We further find that reward design is critical in re-ranking: LLMs often hack rewards by preserving the incoming order or exploiting position bias, motivating conditional verifiable rewards as essential industrial components.

Human or Machine? Turing Tests for Vision and Language

As AI algorithms increasingly participate in daily activities that used to be the sole province of humans, we are inevitably called upon to consider how much machines are really like us. To address this question, we turn to the Turing test and systematically benchmark current AIs in their abilities to imitate humans. We establish a methodology to evaluate humans versus machines in Turing-like tests and systematically evaluate a representative set of selected domains, parameters, and variables. The experiments involved testing 769 human agents, 24 state-of-the-art AI agents, 896 human judges, and 8 AI judges, in 21,570 Turing tests across 6 tasks encompassing vision and language modalities. Surprisingly, the results reveal that current AIs are not far from being able to impersonate human judges across different ages, genders, and educational levels in complex visual and language challenges. In contrast, simple AI judges outperform human judges in distinguishing human answers versus machine answers. The curated large-scale Turing test datasets introduced here and their evaluation metrics provide valuable insights to assess whether an agent is human or not. The proposed formulation to benchmark human imitation ability in current AIs paves a way for the research community to expand Turing tests to other research areas and conditions. All of source code and data are publicly available at https://tinyurl.com/8x8nha7p

Impact of Attention on Adversarial Robustness of Image Classification Models

Adversarial attacks against deep learning models have gained significant attention and recent works have proposed explanations for the existence of adversarial examples and techniques to defend the models against these attacks. Attention in computer vision has been used to incorporate focused learning of important features and has led to improved accuracy. Recently, models with attention mechanisms have been proposed to enhance adversarial robustness. Following this context, this work aims at a general understanding of the impact of attention on adversarial robustness. This work presents a comparative study of adversarial robustness of non-attention and attention based image classification models trained on CIFAR-10, CIFAR-100 and Fashion MNIST datasets under the popular white box and black box attacks. The experimental results show that the robustness of attention based models may be dependent on the datasets used i.e. the number of classes involved in the classification. In contrast to the datasets with less number of classes, attention based models are observed to show better robustness towards classification.