Recommendation & Information Retrieval - Weekly Roundup

Forget bulky atlases and unreliable image searches: MIRAGE offers medical students a free, interactive tool to retrieve, generate, and understand medical images using only open-source models.

OpenSearch-VL offers a fully transparent recipe for training state-of-the-art multimodal search agents, finally democratizing access to a capability previously locked behind closed doors.

Forget scaling laws – the real bottleneck in associative memory isn't storage, it's retrieval: forcing a single "winner" costs you a logarithmic factor in capacity compared to allowing a ranked list.

Forget rigid memory structures: Memini lets your LLM's external knowledge evolve organically, learning and forgetting like a brain.

You can predict EV charging demand surprisingly well using only the first few minutes of a charging session, opening the door to real-time grid optimization.

OpenSearch-VL offers a fully transparent recipe for training state-of-the-art multimodal search agents, finally democratizing access to a capability previously locked behind closed doors.

Forget scaling laws – the real bottleneck in associative memory isn't storage, it's retrieval: forcing a single "winner" costs you a logarithmic factor in capacity compared to allowing a ranked list.

Forget rigid memory structures: Memini lets your LLM's external knowledge evolve organically, learning and forgetting like a brain.

You can predict EV charging demand surprisingly well using only the first few minutes of a charging session, opening the door to real-time grid optimization.

Overcome limitations in capturing complex user-service dependencies with a novel tensor decomposition method that significantly boosts QoS prediction accuracy.

Ditch the attention: ConvRec proves convolutional networks can beat Transformers in sequential recommendation while slashing compute and memory costs.

Forget dumb context stuffing: LongSeeker shows that strategically *editing* its own memory lets agents solve web search tasks with far greater reliability.

Forget relying on LLMs to judge themselves: this "Concept Field" approach uses vector math on text corpora to detect hallucinations and novelty, offering a fast, interpretable, and black-box alternative.

Ditch the vector DB – this new agent architecture achieves SOTA memory recall by storing everything verbatim and optimizing retrieval, all in a single SQLite file.

Your innocent Spotify playlists are leaking surprisingly accurate predictions about your age, habits, and even personality traits, thanks to new AI attack.

Developer-style keyword searches completely nullify the advantage of even the best code embedding models, highlighting a critical gap in current code search techniques.

HeterSEED achieves state-of-the-art performance on heterophilic heterogeneous graphs by decoupling semantic and structural information, offering a more robust approach than relying on feature similarity alone.

TabEmbed leapfrogs existing text embedding models to achieve SOTA performance on tabular data by reformulating tasks as semantic matching problems and using contrastive learning.

E-commerce sentiment analysis is surprisingly influenced by socio-political terminology, impacting the accuracy of customer satisfaction prediction models.

State-of-the-art temporal knowledge graph reasoning is now possible by jointly modeling historical evidence and evolutionary dynamics, unlocking complementary predictive signals.

Political ideology prediction gets a boost: injecting LLMs with knowledge graphs of MP relationships significantly improves accuracy.

RAG systems can be significantly improved by reranking documents based on how much they increase the LLM's confidence, not just their relevance.

Achieve 8x token reduction in million-token document understanding without sacrificing accuracy by having the LLM actively search for relevant information like a foraging animal.

Stop hand-crafting QA datasets for evaluating RAG systems: DoGMaTiQ automates the process with surprisingly high correlation to human judgment, even across languages.

Turns out, chunking code by function is the *worst* way to do retrieval-augmented code completion.

Generative recommenders can slash latency by up to 38% simply by dynamically juggling GPU memory between embedding and KV caches, a feat current systems miss.

Generative recommendation gets a boost: CapsID's soft-routed semantic IDs outperform hard-quantized baselines and even rival sparse-dense hybrids, all while slashing inference latency by nearly half.

Fine-grained analysis of user behavior on search engine results pages is now possible thanks to AllSERP, which adds exhaustive per-element annotations to the AdSERP dataset, covering organic results and widgets in addition to ads.

LLMs for recommendation can now surpass the limitations of static training signals, achieving sustained improvements in ranking accuracy, fairness, and diversity through a dynamically updated Bayesian distillation target.

Forget Gaussian noise - modeling the *decay* of user interest with a custom "burn-down" diffusion process unlocks better personalized recommendations.

On-device LLMs can now drive real-time recommendation improvements, unlocking faster adaptation to evolving user intent without cloud reliance.

Forget bulky atlases and unreliable image searches: MIRAGE offers medical students a free, interactive tool to retrieve, generate, and understand medical images using only open-source models.

Aesthetic quality unlocks better generalization in AI-generated music popularity prediction, beating models trained solely on engagement metrics.

Standard retriever evaluations hide critical weaknesses in agentic search systems, but a new benchmark and training method exposes and addresses these flaws.

LLMs struggle to navigate the complex, multi-turn justification and response dynamics of real-world patent examination, revealing critical gaps in legal reasoning and technical novelty judgment.

Neural retrievers, despite their success on standard benchmarks, fail spectacularly when forced to reason about set-theoretic constraints, revealing a reliance on spurious correlations rather than true compositional understanding.

Learn to build and evaluate your own NLP pipeline, from tokenization to RLHF, using open-weight models and reproducible research practices.

Online advertising can harm users not just through unequal distribution of opportunities, but also by systematically depriving certain groups of relevant concepts or saturating them with skewed framings.

Retrieval-augmented in-context learning, despite its benefits, leaks surprising amounts of private data, even when attackers only have access to paraphrased queries.

Retrieval-augmented LLMs are surprisingly vulnerable to memory poisoning via synonym substitution, a loophole that gradient-based defenses can't close.

Storage scarcity in edge caching doesn't just impact performance, it fundamentally shifts the economic landscape, amplifying inequality among content providers.

GPT-5, combined with physics-based tools, can match traditional scoring functions in ranking protein-ligand docking poses, opening avenues for interpretable curation in drug design.

Fine-tuning dense retrievers on a mix of domain-specific and general question-answering data achieves surprisingly robust performance across diverse legal search tasks, outperforming models trained solely on legal data.

LLMs can now directly generate relevant Point-of-Interest (POI) candidates for map search by encoding both semantic and geographic context, outperforming traditional retrieval methods.

LLMs alone can't capture the nuances of mathematical research, but injecting aspect-aware information into a heterogeneous GNN unlocks surprisingly effective paper recommendations.

RAG systems can now reduce unsafe answers by 37% using SURE-RAG, a transparent evidence verification method that outperforms even GPT-4o in controlled sufficiency tasks.

RAG's reputation for being ineffective in reasoning tasks is shattered by showing that retrieving the right data – intermediate "thinking traces" – unlocks substantial performance gains, even for state-of-the-art models.

Ditch the brittle RAG stack: a unified PostgreSQL data layer slashes latency by up to 92% and eliminates data leakage, making production RAG finally reliable.

Manual student-to-project matching is dead: TeamUp forms better, more diverse teams at scale for pennies per student.

Forget sifting through walls of text – now you can pinpoint exactly where the AI found its answer, down to the pixel, even in complex visuals like charts and diagrams.

Tree-based RAG gets a major upgrade: $\Psi$-RAG's adaptive hierarchical index and multi-granular retrieval agent leapfrog existing methods on complex, cross-document reasoning tasks.

LLMs can now generate 70% syntactically correct and geometrically consistent 3D objects from text, thanks to retrieval-augmented code synthesis.