ZeBROD is a zero‑retraining framework that tackles catastrophic forgetting in object detection by combining YOLO11n for localization with DeIT and Proxy Anchor Loss for feature extraction. It classifies products using cosine similarity against embeddings stored in a Qdrant vector database, enabling accurate detection of both new and existing items without retraining. In a retail store experiment with 140 products, ZeBROD achieved high accuracy and nearly three times faster training than traditional methods, while maintaining an inference time of 580 ms per image on an edge device.
By Priyanto Hidayatullah, Nurjannah Syakrani, Yudi Widhiyasana, Muhammad Rizqi Sholahuddin, Refdinal Tubagus, Zahri Al Adzani Hidayat, Hanri Fajar Ramadhan, Dafa Alfarizki Pratama, Farhan Muhammad Yasin
The paper introduces fixed universal transformers, which are transformers with immutable internal parameters that can emulate any transformer within a specified class by encoding the target model’s description into the input embedding. The authors provide explicit sparse constructions that achieve universality when the embedding dimension is large enough, and demonstrate that universality is generic—randomly initialized transformers are almost surely universal. Empirical tests on parenthesis balancing and multi‑hop reasoning tasks support the theory, suggesting that a transformer’s expressive power largely stems from its input representation rather than its learned weights.
By Jingwen Liu, Alexandr Andoni, Daniel Hsu
HyperFuse is a new, label‑free pipeline for fast self‑supervised node embeddings on attributed hypergraphs. It computes structural node coordinates via a spectral relaxation of hypergraph modularity, builds multi‑scale feature summaries with utility‑weighted hyperedges, and trains a lightweight encoder for only 100 epochs. In experiments on nine public hypergraphs, HyperFuse achieved 13–179× speed‑ups over baselines and matched or exceeded their accuracy on most downstream tasks.
By Megha P, Harshit Kumar, Srajan Agarwal, Anirban Banerjee, Olaf Wolkenhauer, Saptarshi Bej
The study explores how the language used for reasoning affects retrieval‑augmented generation (RAG) in a monolingual German setting. Using a German RAG question‑answering testbed based on the tabletop game The Dark Eye, the authors show that aligning the reasoning language with the query and retrieved documents improves performance, with German reasoning outperforming French reasoning. However, German reasoning still does not surpass the model’s native English reasoning, indicating that native multilingual reasoning is necessary for optimal results.
By Oliver Hauck, Mario Sanz-Guerrero, Katharina von der Wense
The paper introduces CourseChat, an on‑premises, multi‑course retrieval‑augmented generation (RAG) tutor designed for undergraduate business education. It runs behind a campus web gateway, with each of six courses identified by a unique course reference number (CRN) sharing dual AI hosts that provide a FastAPI service, a local vector database, and a local large language model (LLM) served by Ollama. The authors evaluate different model sizes, noting that larger models failed speed requirements while a 12B and 7B model met the classroom speed gate; a mixture‑of‑experts variant improved some corrections but introduced new errors, leading them to retain an 8B model for production pending further improvement. The study highlights that model selection, evidence sourcing, serving compatibility, and product design must be considered together, though it does not demonstrate learning gains and calls for separate evaluation of faculty ratings, peak‑load capacity, and public‑gateway acceptance.
By Sidney Shapiro, Joshua Lindemann
CLIMB is a training‑free inference‑time framework for multimodal retrieval‑augmented generation. It builds a compact complementary evidence pool using an MMR‑style objective that balances relevance and redundancy, then refines answers with a confidence‑controlled critic that scores relevance, specificity, and cross‑modal alignment. The method stops refinement when confidence no longer rises, improving performance on Encyclopedic‑VQA and InfoSeek without altering the retriever or language model.
By Hang Gao, Wujiang Xu, Zhixing Zhang, Kai Mei, Jingyi Yang, Dimitris N. Metaxas
The paper introduces Sigma, a large-scale continuous diffusion language model (3B/8B parameters) that uses steerable, low-dimensional ODE/SDE latent trajectories to address the non-smoothness of discrete diffusion models. Sigma is trained blockwise via likelihood optimization, jointly denoises Gaussian-corrupted token embeddings, and learns an optimal embedding geometry, leveraging pre-trained autoregressive weights for faster training. During inference, classifier-free guidance and score temperature are identified as essential for high-fidelity reasoning and coding, and Sigma matches or exceeds discrete models on benchmarks such as GSM8K, Minerva, HumanEval, MBPP, MATH-500, and AIME, while also revealing unique structural benefits like embedding-space steering and graceful degradation for low NFEs.
By Zhihan Yang, Wei Guo, Jean-Marie Lemercier, Simon Welker, Yonggan Fu, Mohammad Mahdi Kamani, Sajad Norouzi, Julius Berner, Tomas Geffner, Karsten Kreis, Yongxin Chen, Molei Tao, John Thickstun, Pavlo Molchanov, Ante Juki\'c, Arash Vahdat, Morteza Mardani
The paper introduces Adaptive Mask Selection and Composition (AMSC), a method that estimates task similarity using non‑parametric Wasserstein embeddings from state‑action‑reward samples. AMSC selects and weights previously learned policies based on z‑score‑normalized sparsemax similarity scores, forming a prior for new tasks. Experiments on CT‑graph, MiniGrid, and Continual World show that AMSC improves mean performance and forward transfer while avoiding forgetting, highlighting the importance of selecting relevant prior knowledge and its layer‑specific composition.
By Saptarshi Nath, Inish M. D'Souza, Antonio Carta, Soheil Kolouri, Andrea Soltoggio
The paper evaluates how robust large language models (LLMs) are when using retrieval‑augmented generation (RAG) in practical settings. It investigates whether RAG always outperforms non‑RAG approaches, whether adding more retrieved documents helps, and whether the order of documents matters, using a benchmark of 1,891 samples across five datasets and three task categories. Experiments with 11 LLMs show generally high retrieval robustness, but performance varies by task and prompting strategy, indicating that adopting RAG should be considered on a case‑by‑case basis.
By Shuyang Cao, Karthik Radhakrishnan, David Rosenberg, Steven Lu, Pengxiang Cheng, Lu Wang, Shiyue Zhang
The paper introduces a hierarchical model-based reinforcement learning framework that learns control policies from action-free time series by exploiting shared dynamics across related systems. It uses low-dimensional dynamical embeddings to capture both shared structure and individual variation, which then parameterize shared policy and value networks. Experiments on Lorenz-63, double-pendulum, and neural-behavioral data show that these hierarchical policies outperform independently trained ones, achieve higher rewards than planning with the same models, and generalize to unseen systems using only embedding inference.
By Niklas Emonds, Georgia Koppe
Cephalonauts One is a 3 Tesla fMRI dataset featuring 30 hours of brain activity recorded from three healthy subjects while they listened to native-language audio podcasts. The release includes the raw fMRI data, corresponding podcast audio, transcript annotations, and derived stimulus embeddings, making it the deepest naturalistic speech fMRI dataset available. A brain‑decoding benchmark is introduced, framing audio segment retrieval as a task where decoders must match fMRI activity to the correct time‑aligned podcast segment, with standardized splits, metrics, and baseline models provided.
By Antoine Collas, Louis Jalouzot, G\'eraud Ilinca, Corentin Caris, Romain Valabr\`egue, Ahmed Hassayoune, David Goncalves, Madeleine Hueber, Thadd\'ee Delebarre, Julien Savatovsky, Clara Fonteneau, Charles Maussion, Bertrand Thirion, Alexis Thual
FASTDIAR is a frame‑level speaker encoder designed for real‑time streaming diarization on a single CPU thread. It converts a state‑of‑the‑art speaker recognition architecture into a causal model that emits an embedding every 80 ms from a two‑second window, and pairs it with online clustering that updates only when the stream’s self‑similarity changes. Trained via distillation from an utterance‑level teacher, it achieves the highest accuracy on low‑overlap benchmarks, degrades less with more speakers, and runs five times faster than real time.
By Nikita Torgashov, Okan K\"op\"ukl\"u
The paper investigates the geometric capacity of vector retrieval systems, focusing on the maximum recall achievable with a fixed document index. It demonstrates that, on real-world benchmarks, single-vector query encoders often underperform relative to the index’s potential. The authors provide theoretical evidence that learning such encoders can be computationally hard, constructing a task where a simple neural network can achieve perfect recall while any statistical-query learner would need exponentially many queries to surpass random chance.
By Anders Wikum, Nina Mishra, Amin Saberi, Tal Wagner
The paper introduces Calibrated User Embeddings (CUE), a framework that encodes real user sessions into continuous representations and decodes them into persona commands to steer large language models (LLMs) as user simulators without additional training. CUE enables user-conditioned replay of past interactions and generates novel personas that better align with real-user success rates and failure patterns. Evaluations on the τ²‑Bench dataset show that CUE‑driven simulators commit fewer simulator‑attributed errors, more accurately reproduce real‑user failure modes, and maintain competitive user fidelity across various tasks and LLMs.
By Anjali Kantharuban, Jonas Mueller
The paper introduces a controlled anchor–stress protocol to analyze compositional failures in text‑to‑image diffusion models. It presents a text‑only Compositional Stress Index (CSI) that distinguishes common from rare compositions across SD1.5, SDXL, and SD3, and conducts a six‑prompt localization study showing how different denoiser interventions affect outcomes such as residual minimization and color hit rate. The study reveals that diagnostic signals appear before denoising, but the intervention that exposes risk is not always the one that improves generation, highlighting modality‑ and architecture‑dependent responses.
By Fangzheng Wu, Brian Summa
S2S-JEPA is a new AI weather model that focuses on predicting only the slowly varying components of the atmosphere at subseasonal-to-seasonal timescales, from two weeks to two months ahead. It adapts the Joint-Embedding Predictive Architecture (JEPA) from computer vision to discard unpredictable fine-scale details, thereby addressing the long‑known predictability desert. The model matches the skill of the ECMWF physics‑based ensemble and even outperforms it on several metrics during weeks 5 to 6.
By Chenyu Dong, Gianmarco Mengaldo
The paper presents a method for improving cross‑lingual retrieval in multilingual encoders by adding a query‑only low‑rank adapter and deterministic routing based on language. The adapter, trained against frozen document embeddings, is used for cross‑lingual queries while the original encoder handles same‑language queries, preserving existing performance. Using the SampoTron LoRA adapter with the Nemotron‑3‑Embed‑1B model, the authors report a 20.9% relative gain in nDCG@10 across six English, Finnish, and Swedish directions on a financial benchmark, with all cross‑lingual directions improving and same‑language performance maintained.
By Akshay Jain, Edward Kim
AIBL (Augmented Instance-Based Learning) extends traditional Instance-Based Learning Theory by embedding high‑dimensional sequential data into a learned vector space, enabling neural similarity matching while preserving instance storage, activation‑weighted retrieval, and utility blending. The model introduces structured memory stores—active, forgotten, and surprise—to manage novelty detection and concept drift, promoting frequently surprising instances into active memory. Experiments on five machine‑learning tasks and three simulations show that AIBL outperforms classical IBLT variants and task‑specific baselines, improving accuracy by 6 to 17 percentage points.
By Radha Poovendran, Andrea Stocco, Linda Bushnell
The study applies contrastive representation learning (CEBRA) to EEG data collected from conversational dyads, embedding the recordings onto a 2‑dimensional sphere. The resulting manifold captures individual identity features, as evidenced by high decoding accuracy for dyad labels and significant differences when labels are permuted only after retraining the encoder. However, the embedding shows little sensitivity to conversational role (speaker vs. listener) and to autism‑quotient differences, suggesting that the geometry is dominated by individual traits rather than task‑specific dynamics.
By Hubert Huang, Michelle McCleod, Brendan Ames, Evie Malaia
The paper presents a system for matching job candidates to vacancies that provides interpretable evidence rather than a single relevance score. It uses a two‑stage approach: an LLM‑based labeler refined through recruiter feedback and a distilled bi‑encoder that runs online on CPU. The model, trained on 168,772 labeled pairs, achieves 95.79% agreement with recruiter‑recorded decisions on a production‑feedback subset.
By Ilya Chekin (BroutonLab), Vyacheslav Malyugin (BroutonLab), Vladimir Chirkov (BroutonLab), Mikhail Yurushkin (Curately)