Paper Daily · Generative (English)

PailitaoMMSearch2.0 Technical Report: Scaling Generative Retrieval to Unified End-to-End Multimodal Search

Thu, 29 Oct 2026 08:00:00 +0800

In Taobao's Pailitao, a user may photograph a shirt and ask for shorts to match it, which calls for a complementary item rather than a duplicate. The usual pipeline classifies intent first and routes to separate retrieval channels, so one wrong step breaks everything after it. Alibaba's report lets a 2B vision-language model read the raw image and instruction and generate product Semantic IDs dir…

Generative · Score 4.50 · Deep dive · 36 min · Xiaohan Ye, Zihan Gong, Jingqian Zhao, Lianyu Du, Yunmeng Shu, Xu Chen, Zesheng Zhou, Renchun You,… · Source

KL-MOPD: Discrepancy-Aware Multi-Objective On-Policy Distillation for Generative Recommendation

Thu, 29 Oct 2026 08:00:00 +0800

When click and value teachers are distilled together into a generative recommender, equal per-token weights are unfair. On TencentGR the student's KL to the value teacher at the first token is 3.13 times its KL to the click teacher, and the harder teacher switches to the click teacher on the last two tokens. KL-MOPD from Zhejiang University and Alibaba weights each teacher per token by its pre-tr…

Generative · Score 4.20 · Deep dive · 25 min · Weiqin Yang, Yandong Wang, Yuan Gao, Hao Deng, Jinxin Hu, Jing Cao, Jiawei Chen · Source

Latent Residual Reasoning for Ranking in Generative Recommendation

Thu, 29 Oct 2026 08:00:00 +0800

Generative ranking models usually score every candidate in one forward pass. Refiner, from Meituan and the University of Science and Technology of China, adds a few more steps. Each candidate gets a reasoning token, and each step predicts only a correction added to the original score. The base model stays frozen; only LoRA and a new output layer are trained. On Steam it lifts MTGR's AUC from 0.76…

Generative · Score 4.20 · Deep dive · 34 min · Haihan Shi, Ruidong Han, Wenjie Wang, Shangyu Chen, Sunhao Dai, Yongqi Li, Yangyang Li, Fei Jiang,… · Source

REGER: Modeling inter-item Relations with Semantic IDs for Generative Recommendation

Thu, 29 Oct 2026 08:00:00 +0800

Cross-entropy tells a generative recommender only which token is correct, with no signal about how other candidates compare. REGER, from the University of Electronic Science and Technology of China, splits a conventional sequential model's scores over all items into branch preferences along the SID prefix tree and distills them along the true path. Added alone to TIGER, it lifts Beauty Recall@10…

Generative · Score 4.10 · Deep dive · 27 min · Yuxi Huang, Silin Zhou, Chenhao Wang, Zihan Li, Gangming Guo, Pei Ke, Shen Gao, Shuo Shang · Source

LESSRec: Level-Specific History Modeling for Semantic ID-Based Recommendation

Thu, 29 Oct 2026 08:00:00 +0800

When a semantic-ID recommender predicts a token at one level, it attends to historical tokens at every level. LESSRec's inference-time analysis finds that history at the target token's own level is more informative and other levels can interfere. So it encodes history separately per level, attends only to the matching level during prediction, and adds an item-level semantic sequence, fusing both…

Generative · Score 4.10 · Shangxin Yang, Zongwei Wang, Jing Long, Junliang Yu, Min Gao · Source

JevRec: Linguistic Encoding with Jev-like Decision Modeling for LLM-based Generative Recommendation

Thu, 29 Oct 2026 08:00:00 +0800

Most generative recommenders give every item an SID of the same length, so popular items keep paying for fine-grained codes, while shorter codes squeeze long-tail semantics. JevRec borrows the natural-language pattern that frequent words are short: popular items get short SIDs and long-tail items long ones. A dedicated head predicts the length first and a separate head generates content condition…

Generative · Score 4.10 · Deep dive · 30 min · Qiyong Zhong, Jiajie Su, Ming Yang, Li Zhang · Source

Modeling Preference Dependencies with Directed Acyclic Graphs for Narrative-Driven Recommendation

Thu, 29 Oct 2026 08:00:00 +0800

When users describe needs in free text, preferences often qualify one another, as in a seaside city that suits elderly parents but is not too far away. Existing methods treat the text as flat keywords or hand it to an LLM to generate queries freely. PDG-F first has an agent parse the request into a directed acyclic preference dependency graph, then retrieves candidates from the open web along gra…

Generative · Score 3.85 · Mingzhe Lu, Jian Qin, Ruize Qin, Jinhao Liu, Yunpeng Li, Hang Su · Source

RPSD: Retrieval-Privileged On-Policy Self-Distillation with Optimal Transport

Thu, 29 Oct 2026 08:00:00 +0800

When an LLM recommender is post-trained with GRPO, one sequence-level reward is shared by every token, and each signal comes from a single user. RPSD has a teacher with privileged information re-evaluate the student's own response, and uses their per-token confidence gap to turn the sequence reward into token-level credit. The privileged information comes from similar users: time-aware optimal tr…

Generative · Score 3.75 · Dongyue Li, Zebang Cheng, Zhongyi Deng, Hongboxu, Yi Bin, Fei Ma · Source

The Vocabulary Is Already Here: Adapting Pretrained Interpretable Features for Generative Recommendation

Thu, 29 Oct 2026 08:00:00 +0800

Existing generative recommenders train a dedicated SID vocabulary fitted to language-model representations of item descriptions. This paper notes that pretrained sparse autoencoders already expose language-model features with human-readable annotations, so feature indices can serve directly as the SID vocabulary. The API framework keeps each item's selected feature indices and relative activation…

Generative · Score 3.75 · Ziluowen Luo, Ruochen Liu, Jun Yin, Hao Yan, Zhengxuan Hu, Xinze Shi, Chaozhuo Li, Chengqi Zhang, S… · Source

MAREval: A Multi-Agent Framework for Evaluating Natural Language Recommendation Explanations

Thu, 29 Oct 2026 08:00:00 +0800

Judging recommendation explanations by hand is accurate but expensive, while a single LLM judge is prompt-sensitive and inconsistent. MAREval splits the job across agents: a planner assigns roles with debate-style prompting, a moderator keeps the discussion on track, an arbitrator aggregates multiple rounds, and Monte Carlo sampling adds robustness. On a public benchmark and an e-commerce dataset…

Generative · Score 3.65 · Reza Yousefi Maragheh, Jayesh Uddhav Kudase, Aysenur Inan, Ramin Giahi, Kai Zhao, Parth Parikh, Jia… · Source

Octen-WebGR: Scaling Generative Retrieval to 200 Million Web Pages

Wed, 28 Oct 2026 08:00:00 +0800

Generative retrieval has stayed below ten million documents. The Octen team's Octen-WebGR (a model that does web navigational search with generative retrieval) pushes it to 201 million web pages: every page is named by two tokens, a 4B model generates them directly from the query, and a table lookup returns the URL. On real navigational traffic from the month after training, it ranks the target f…

Generative · Score 4.40 · Deep dive · 30 min · Guangwei Xu, Wenkai Zhang, Yawen Liu, Dongxu Zhang, Xijie Xu, Kuan Zou · Source

PinGER: Token-Efficient Generative Retrieval with Semantic IDs

Wed, 28 Oct 2026 08:00:00 +0800

Generative retrieval encodes each item as a multi-level SID and generates it level by level, but on the input side every history item costs several tokens. During training, Pinterest's PinGER places one dense-vector token after each item's SID and requires that any token reading another item may see only that item's dense token, never its SID. At serving time all SIDs in the history can be droppe…

Generative · Score 4.35 · Deep dive · 38 min · Xinyi Li, Sujan Reddy Ale, Yi-Ping Hsu, Zelun Wang, Yilin Chen, Hanyu Li, Hanlin Lu, Arkin Dharawat… · Source

SLIDE: Search-to-List Item Distillation for Single-Rollout Generative Retrieval

Wed, 28 Oct 2026 08:00:00 +0800

Generative retrieval usually produces candidates online with beam search, and the beams never see each other; writing a whole list in one go by greedy decoding is the most direct alternative but works much worse. Taobao's SLIDE (a training method that distills offline search results into lists) moves search offline: a teacher runs beam search for every training request, concatenates the retrieved…

Generative · Score 4.30 · Deep dive · 34 min · Tianyu Zhan, Gui Ling, Ran Tao, Kunhai Lin, Tong Xiong, Wang Yang, Kaixuan Zhang, Zhihong Chen, Yul… · Source

MuSID-KD: Multigranularity Knowledge Distillation for Generative Recommendation with Semantic IDs

Wed, 28 Oct 2026 08:00:00 +0800

When a SID-based generative recommender is distilled into a small model, the student is usually asked only to match the teacher's distribution over which item to recommend in the end. ECNU's MuSID-KD (a multi-granularity distillation method that works level by level over the SID) finds this is not enough: a student aligned only on items disagrees with the teacher on the order of candidate groups…

Generative · Score 4.10 · Deep dive · 32 min · Fengshi Zhang, Xiyan Ding, Shuxiao Xie, Wei Zhang · Source

LoopGR: Parameter-Efficient Scaling for Generative Recommendation via Recurrent Depth

Wed, 28 Oct 2026 08:00:00 +0800

Making a generative recommender stronger usually means adding layers, which means adding parameters and memory. Tencent's LoopGR (a generative recommendation backbone that reuses one decoder in a loop) runs the same decoder several times. Between passes, the decoder's BOS hidden state (its current judgement of the target) is added to every history position and a tiny GRU rewrites the history alon…

Generative · Score 4.00 · Deep dive · 30 min · Huizhe Zhang, Yudong Li, Liang Chen, Chengxiang Zhuo, Zang Li · Source

DPE-Rec: Diversity-Preserving and Efficient Diffusion Language Model Recommendation for Bundle Construction

Wed, 28 Oct 2026 08:00:00 +0800

Completing a product bundle or playlist is unordered generation of a set of items, so a diffusion language model with bidirectional attention suits it, and it can reveal several tokens per step. Each item is a multi-level SID whose first token decides the coarse category, but with confidence-ordered revealing the model postpones that token, so generated items look alike; revealing too many at onc…

Generative · Score 3.80 · Pingchuan Ma, Bohan Jiang, Chengshuai Zhao, Dawei Li, Zhen Tan, Huan Liu · Source

ReflectRec: Cycle-Consistency Preference Verification for LLM-based Recommendation

Wed, 28 Oct 2026 08:00:00 +0800

Chain-of-thought LLM recommenders first write a user preference note and then rank, but the note is never checked and the ranker never consults behaviorally similar users. ReflectRec (an LLM recommendation method that verifies preference notes by a round trip) has the model guess which movies the user watched from the note alone, and uses the share of the real history it recovers as the note's tr…

Generative · Score 3.65 · Dae Kyoo Kim, Heejung Cho, Sungmoo Hong, Eunil Park · Source

Transporting Quantiles to Codebook: Scalable Vector Quantization without Codebook Collapse

Wed, 28 Oct 2026 08:00:00 +0800

Vector quantization underlies image and audio generation and even recommendation SIDs, but the common VQ-VAE codebook often collapses, leaving most codewords unused. FlowVQ (a flow-matching-based vector quantization method) first cuts a fixed simple prior distribution into cells of equal probability, one initial codeword per cell. A lightweight flow model trained together with the VQ-VAE then map…

Generative · Score 3.45 · Yuwei Zeng, Zekun Shi · Source

PinRec: Unified Generative Retrieval Model for Pinterest Recommender Systems

Tue, 27 Oct 2026 08:00:00 +0800

Recommender systems usually run a separate retrieval model for the home feed, search and related items, because each surface wants something different: saves on the home feed, outbound clicks to buy on search, and continued browsing on related items. This Pinterest KDD 2026 paper replaces all three with one generative retrieval model, PinRec. It is pretrained on site-wide behavior, fine-tuned on…

Generative · Score 4.10 · Deep dive · 34 min · Edoardo Botta, Jaewon Yang, Yi-Ping Hsu, Laksh Bhasin, Yilin Chen · Source

Unifying Behavior Modeling and Semantic Generation for Generative Recommendation

Tue, 27 Oct 2026 08:00:00 +0800

Generative recommenders that write the next item's Semantic ID code by code have a systematic bias. This KDD 2026 paper from South China University of Technology finds that the fewer items in a user's history share the target's category, the worse they do against classic behavior-only sequential models; in the worst group, TIGER's NDCG@10 trails SASRec by 30%. The cause: the encoder learns which…

Generative · Score 4.00 · Deep dive · 38 min · Binquan Wu, Xinbo Chen, Yicheng Luo, Yuhao Ke, Jingye Li · Source

CoFiRec: Coarse-to-Fine Tokenization for Generative Recommendation

Tue, 27 Oct 2026 08:00:00 +0800

When shopping online, people usually look at a broad category before picking a specific item. Yet most generative recommenders build Semantic IDs by mashing an item's category, title, description and collaborative signal into one vector before quantizing it, so that coarse-to-fine structure is lost. This UIUC and Meta paper proposes CoFiRec, which splits an item into category, text and collaborat…

Generative · Score 4.00 · Deep dive · 18 min · Tianxin Wei, Xuying Ning, Xuxing Chen, Ruizhong Qiu, Yupeng Hou · Source

Separating Thinking from Spelling for Efficient Generative Recommendation

Tue, 27 Oct 2026 08:00:00 +0800

Generative recommenders retrieve by writing an item's Semantic ID code by code, and every beam-search step runs through the large backbone, so the cost grows with more candidates and longer codes. It is often the main expense. This paper proposes Deco, which separates reading the history from writing the code. The backbone reads the history once and condenses it into a vector for the next item, a…

Generative · Score 3.90 · Danil Gusak, Anna Volodkevich, Ivan Oseledets, Evgeny Frolov · Source

Knowledge Distillation across Item and Code Spaces for Parallel Generative Recommendation

Tue, 27 Oct 2026 08:00:00 +0800

In generative recommendation with Semantic IDs, training only tells the model which item came next, so it learns nothing about how likely the other items were given the same history. A classic sequential recommender can provide that distribution, but it scores items while the generative model predicts codes position by position, so the knowledge is hard to distill directly. This paper from HKUST…

Generative · Score 3.90 · Deep dive · 23 min · Yixuan Xu, Hailin Liu, Linxiang Peng, Can Yang · Source

GUIDER: Generative User Interest Discovery via Explicit Reasoning with Large Language Models

Tue, 27 Oct 2026 08:00:00 +0800

The Semantic IDs common in generative recommendation are opaque strings of codes: it is hard to see what the model generated or why, and the LLM's own reasoning goes unused. GUIDER splits recommendation into two steps. The LLM first reasons in text about which interest circles the user may want next. Circles are interest topics described in words and must come from a fixed set; items are then mat…

Generative · Score 3.75 · Jinke Wu, Yinghao Wu, Shuchang Liu, Yandong Bai, Qi Zhu · Source

SAGE: Sequence-level Adaptive Gradient Evolution for Generative Recommendation

Tue, 27 Oct 2026 08:00:00 +0800

Generative recommenders are often trained with supervision first, then aligned to user feedback on whole slates with reinforcement learning. To keep training stable, each update may change the probability of generating a slate only by a fixed ratio set per sample, which smothers rare positive feedback on cold items and barely punishes repetitive negative slates. This Xiaohongshu paper proposes SA…

Generative · Score 3.75 · Yu Xie, Xing Kai Ren, Ying Qi, Ping Yang, Yao Hu · Source

TGR-EvoDAG: A DAG-based Multi-Agent Code Evolution Framework for Generative Recommendation on Tencent

Tue, 27 Oct 2026 08:00:00 +0800

Letting LLM agents rewrite code to search for better recommenders runs into two problems in generative recommendation. Retraining the tokenization, pretraining and alignment stages costs very different amounts, yet agents spend budget as if it were uniform. And a good upstream change often makes metrics worse until downstream stages adapt, so it gets discarded too early. This Tencent paper propos…

Generative · Score 3.65 · Junlin He, Junzhang Zhu, Yudong Li, Yihong Tang, Xuefeng Jiang, Wei Ma, Chengxiang Zhuo, Zang Li · Source

From Item Scores to Semantic Tokens: Test-Time Collaborative Adaptation for Generative Recommendation

Mon, 26 Oct 2026 08:00:00 +0800

In generative recommendation with semantic IDs, the model writes an item's code one token at a time, and beam search keeps only the highest-scoring prefixes at each step. Once a prefix is pruned, every item beneath it can never be generated. The problem is sharpest when a user has just developed a new interest that shows up in only one or two recent interactions. This paper proposes TCBSA (test-t…

Generative · Score 4.10 · Deep dive · 35 min · Xiongfeng Shan, Xing Tang, Hanchong Chen, Fuyuan Lyu, Lingjie Li, Jingyan Jiang, Duguang Liu, Xiuqi… · Source

MPQ: A Message-Passing View of Post-Training Quantization

Mon, 26 Oct 2026 08:00:00 +0800

When an LLM is quantized after training to very low bit widths, common solvers such as GPTQ decide block by block which codebook value each group of weights takes. Once a block is fixed it never changes, and near-tied alternatives are dropped. This paper proposes MPQ (message-passing quantization), which keeps probabilities over each block's candidate values and commits only at the end. It altern…

Generative · Score 3.00 · Yanshuo Chen, Chenxi Liu, Reza Shirkavand, Heng Huang · Source

How Language Models Learn Taste: Brand Preference in LLMs

Sun, 25 Oct 2026 08:00:00 +0800

Ask LLMs to "name a running shoe brand" and they will most likely say Nike; ask them to "recommend a running shoe brand" and they are more likely to say Brooks. The paper measures this effect across 102 models and 42 categories. In 26 categories the answer changes this way, and the models move together to the same brand: Chrome becomes Firefox, Google becomes DuckDuckGo. The brands named off the…

Generative · Score 4.00 · Deep dive · 32 min · Tapan Parikh · Source

A Cue's Hidden Price: Auditing Seller Influence on AI Shoppers

Sun, 25 Oct 2026 08:00:00 +0800

Two headphone listings are identical in price, star rating and specs, except that one adds the line “Our best-selling model, chosen by over 10,000 customers.” Across six such pairs, GPT-4o picked the listing with that line in all 144 trials. Raise that listing's price step by step, and it has to cost about 13% more before the model starts to waver. With a plain product description instead, the ad…

Generative · Score 4.00 · Deep dive · 31 min · Felix Buecheler, Gerardo Berbeglia, Mingyu Guo · Source

Prompting Strategy Effects Depend on the Judge: A Multilingual Study of LLM-as-a-Judge Robustness

Sun, 25 Oct 2026 08:00:00 +0800

When LLMs act as judges scoring multilingual outputs, people often reuse what works in English, such as asking the judge to write out its reasoning first or to play an expert. The authors tried 9 prompting styles on 4 judge models across 27 languages and found that this experience does not transfer. Which judge model you use matters far more than how you prompt it. The same style can help one jud…

Generative · Score 3.85 · Asal Mehrabi, Mohammad Aliannejadi, Clemencia Siro · Source

LLM Multi-Agent Systems Are Vulnerable to Herding and Runaway Cascades

Sun, 25 Oct 2026 08:00:00 +0800

What happens when a group of LLM agents recommend songs to one another and each can see how many times a song has already been recommended? The authors reran the classic Music Lab experiment from sociology with songs the agents had never heard: one group could see the running recommendation counts, the other could not. The group that could see them quickly piled onto a few songs, and which song b…

Generative · Score 3.75 · Farnoosh Hashemi, Shiyu Ji, Joice Chen, Michael W. Macy · Source

UniRec: A Unified Expressive-Aligned Generative Recommendation Framework for E-commerce

Sat, 24 Oct 2026 08:00:00 +0800

Generative recommendation generates a product's semantic IDs layer by layer, and before the first layer it has seen no product-side information at all; this is its gap with discriminative ranking models. When a discriminative model scores a candidate it sees the product's category, shop and brand directly and crosses them with user features. UniRec has the model first generate the product's categ…

Generative · Score 4.65 · Deep dive · 34 min · Ziliang Wang, Gaoyun Lin, Xuesi Wang, Shaoqiang Liang, Yili Huang, Weijie Bian, Li Zhang, Mingchen… · Source

GenPage: Towards End-to-End Generative Homepage Construction at Netflix

Sat, 24 Oct 2026 08:00:00 +0800

The Netflix homepage used to be assembled by a multi-stage recommender that retrieved candidate rows and titles, ranked them separately and then built the page; GenPage instead generates the whole page with one decoder-only Transformer. The homepage is made of rows, each row a theme. User history, account attributes and request information are encoded as a prompt, and the model generates row toke…

Generative · Score 4.45 · Deep dive · 34 min · Lequn Wang, Jiangwei Pan, Linas Baltrunas · Source

OneLatent: Latent Reasoning for Efficient Foundation Recommendation Models

Sat, 24 Oct 2026 08:00:00 +0800

When a large language model serves as a recommendation foundation model and first writes a natural-language reasoning trace before outputting the next item's semantic ID, the reasoning text is far longer than the semantic ID and holds back both decoding time and throughput. OneLatent appends 3 fixed special tokens after the user context and trains in several steps. First the model learns to repro…

Generative · Score 4.40 · Deep dive · 35 min · Hao Jiang, Peiru Du, Pengfei Yao, Mengting Li, Siyuan Lou, Kuo Cai, Sheng Yu, Qiang Luo, Jian Liang… · Source

Coarse-to-Fine Long-term Interest Modeling for Generative Recommendation

Sat, 24 Oct 2026 08:00:00 +0800

Generative recommendation generates the next item's semantic IDs one code at a time, and attention cost grows quadratically with sequence length, so these models usually take only the most recent few dozen behaviors and leave the long-term history unused. GLASS uses the long-term history differently for each semantic ID level. Before generating the first-level code, it averages the item vectors o…

Generative · Score 4.20 · Deep dive · 28 min · Shiteng Cao, Junda She, Ji Liu, Bin Zeng, Chengcheng Guo, Kuo Cai, Qiang Luo, Ruiming Tang, Han Li,… · Source

CSG4Rec: Collaborative-Similarity Guided Semantic IDs Tokenization for Generative Sequential Recommendation

Sat, 24 Oct 2026 08:00:00 +0800

Generative recommenders encode each product as a sequence of semantic IDs, and this work finds that learning behavioral similarity directly in the space where the IDs are built works better than aligning it afterwards, beating strong existing baselines on three common datasets. Existing methods first build the IDs from product content and then align the ID vectors with embeddings learned by colla…

Generative · Score 3.95 · Wei Liu, Bo Yang, Qianyang Zhu, Ruijie Xiao, Sirui Huang, Qing Li · Source

Evaluating LLM-based conversational recommenders with semantic and genre-aware metrics

Sat, 24 Oct 2026 08:00:00 +0800

When an LLM chats with a user and recommends movies, the usual Recall@K and BLEU only check whether a recommendation is the exact movie the user finally picked in the dataset, yet many recommendations that miss it still suit the user's taste. On ReDial and INSPIRED, two movie-recommendation dialogue datasets, this study found that the reference film appears in the top 10 recommendations only 39%…

Generative · Score 3.85 · Darshna Parmar, Dhruv Shah, Pramit Mazumdar, Saurav Mallik · Source

Integrating implicit desire fusion for proactive conversational recommendation systems with LLMs

Sat, 24 Oct 2026 08:00:00 +0800

In conversational recommendation, first inferring what the user really wants from a few scattered remarks, and only then deciding whether to ask or to recommend, makes success much more likely. On LastFM music recommendation, with GPT-3.5-Turbo talking to a simulated user, adding this step raises the share of sessions that succeed within 15 turns from 0.327 to 0.583. For example, the user says th…

Generative · Score 3.85 · Sitong Yan, Shengli Song, Chengen Lai, Jingyang Li, Guangneng Hu · Source

Controllable Generative Recommendation via Guided Token Refinement