Bert
-
Dynamic ordering in sequence generation in Transformers: experiments on low-resource Tagging scenarios
2026-10-15 13:36:46Giacomo Bais, Tejaswini DeoskarAbstract
This paper introduces TagInsert, a Transformer-based model that challenges the common left-to-right (L2R) generation paradigm in Transformers, for sequence labelling tasks where dependencies may not naturally align with a L2R order. TagInsert lifts the L2R constraint by removing the decoder-side attention mask and enabling the model to decide a preferred order of tag generation at inference time. We experiment on the common Part-of-Speech tagging and the more complex Combinatory Categorial Grammar (CCG) supertagging tasks on multiple languages and low-resource data settings. We show that, although encoder-only models like BERT remain strong performers for tagging tasks, TagInsert frequently outperforms a L2R encoder-decoder model, achieves results comparable to encoder-only models under data-scarce conditions, and shows better results on the long-tail of rare categories. Our analyses show that TagInsert alleviates error propagation typical of autoregressive models by learning to defer difficult predictions to later stages in the decoding process.
Abstract (translated)
URL
https://arxiv.org/abs/2610.21862
PDF
-
SITAR: A Situational Distortion Metric for Automatic Speech Recognition Evaluation
2026-10-15 07:48:03Mingzhe Lu, Jinhao Liu, Yunpeng Li, Hang Su, Lichun Fan, Jian LuanAbstract
In automatic speech recognition (ASR), transcription errors with similar lexical or semantic discrepancies can have markedly different consequences for the situation conveyed by speech. Capturing these differences requires not merely contextual assessment, but an understanding of what the spoken discourse describes, establishes, or seeks to bring about. However, relying on LLM judges introduces computational cost and scoring variability. We introduce SITAR, a situational distortion metric for ASR, realized through a comprehensive end-to-end construction framework. Specifically, this framework aggregates situational severity and cumulative extent into a unified score via a bounded recursive formulation. Using targets derived from this formulation with limited offline LLM support, we train SITAR as a lightweight BERT-family regressor that scores reference-hypothesis pairs in a single forward pass, entirely free of any prohibitive runtime costs. Recognizing that introducing a new metric demands rigorous justification, we formalize a principled validation protocol spanning both extrinsic validity, assessed at macro and micro levels, and intrinsic scorer reliability. The former, evaluated across diverse datasets and models, demonstrates that SITAR reliably tracks broad capability scaling while providing orthogonal diagnostic utility, whereas the latter, examined on a purpose-built controlled benchmark, confirms the operational soundness of the scoring instrument itself. Through systematic validation, we demonstrate that SITAR holds strong potential as a practical, lightweight complement to existing ASR metrics. We will release the full recipes (code, model weights, dataset, etc.) after the internal review to support community research.
Abstract (translated)
URL
https://arxiv.org/abs/2610.21440
PDF
-
Code-Switching Detection via Language-Selective Dimensions in Speech Foundation Models
2026-10-14 22:26:33Jie Chi, Zakaria Aldeneh, Masha Fedzechkina, Maureen de Seyssel, Natalie SchluterAbstract
Code-switching (CS), the alternation between languages within an utterance, is common in multilingual speech but underrepresented in pretraining corpora. Increasing CS coverage requires finding relevant utterances in large speech collections so that they can be included in training data. One way to find such utterances is to transcribe the speech and detect language alternation in the text, but this depends on obtaining accurate multilingual transcriptions. We propose a lightweight method that detects CS directly from speech. It uses frame-level language information in pretrained speech encoders to determine whether both languages are present within an utterance. Specifically, we identify language-selective hidden-state dimensions using monolingual recordings and aggregate their frame-level activations into a CS prediction. With a small CS set used to tune a decision rule, we show that the Whisper-small encoder achieves a mean F1 of $0.798$ across 14 language pairs in CS-FLEURS, compared with $0.776$ for a CNN and $0.751$ for a HuBERT linear baseline trained for monolingual language identification. Our findings suggest that pretrained speech encoders already encode the information needed to identify CS utterances, and they provide a lightweight tool for building more representative multilingual corpora.
Abstract (translated)
URL
https://arxiv.org/abs/2610.18726
PDF
-
Calibration-Preserving Pruning: Compression as a Reliability Contract
2026-10-14 18:40:25Ibne Farabi Shihab, Adria Binte Habib, Joyanta Jyoti MondalAbstract
Split conformal prediction, not the pruning rule, supplies finite-sample marginal coverage once a pruned model is fixed independently of the conformal calibration split. We study the separate efficiency problem: can pruning preserve score geometry well enough to obtain smaller valid prediction sets? Calibration-Preserving Pruning (CPP) augments a base pruning score with nonconformity-gradient saliency and uses disjoint pruning, validation-selection, conformal-calibration, and test splits. Bounded score perturbations imply bounded conformal-quantile shifts and controlled set inflation, but do not make the generic coverage theorem CPP-specific. Final five-seed Qwen2.5-1.5B results at 50\% sparsity show the largest gains on large-label tasks. On DBpedia-14, CPP-SparseGPT reduces mean set size from \(10.1\) to \(8.6\) while changing accuracy from \(0.347\) to \(0.366\); CPP-Wanda reduces \(11.2\) to \(9.0\) with an accuracy trade-off from \(0.310\) to \(0.295\). Across 15 dataset--sparsity cells, CPP-SparseGPT produces smaller sets in 13 and higher accuracy in 11. Matched controls show that generic supervised gradients explain much of the gain: true-label CPP is not statistically resolved from matched Wanda+SNIP, whereas threshold-aware candidate-label CPP reaches \(7.8\) mean set size at explicit accuracy and offline-compute costs. RoBERTa-base and Llama-3-8B diagnostics support transfer, but our claims remain limited to reliability-sensitive classification.
Abstract (translated)
URL
https://arxiv.org/abs/2608.23744
PDF
-
Reliability-Aware Multimodal Sentiment Prediction Under Local Modality Missingness
2026-10-14 17:36:18Minghao Kong, Deng Yao, Xi Zhang, Ying Gao, Jiurun Chen, Zirong Liu, Rongjie WangAbstract
Multimodal sentiment models are usually evaluated on complete utterances, yet streams often lose a contiguous span of text, audio, or video. We study local missingness on aligned CMU-MOSEI features. Our reliability-aware network (RM-Net) keeps the original temporal positions, represents unobserved positions with a learned token, outputs bounded per-modality intensity contributions gated by observation quality, and is trained with dual-view masking of contiguous segments. We fuse five RM-Net seeds with a LoRA-adapted BERT text expert, using a fusion weight that depends only on the visible fraction of content tokens. Architecture, checkpoint, and fusion choices were made on the validation split. On the held-out split, the fused model reaches a macro-F1 of 0.655 on clean inputs and 0.634 under 40\% mixed missingness; its two-view average exceeds the auxiliary-head baseline by 0.024 (95\% video-cluster bootstrap interval [0.003, 0.045]). Because this split was inspected during development, these numbers are not a blind test. For explanations, exact Shapley attributions over the eight modality subsets and position-level occlusion show that top-ranked positions reduce the predicted-class probability faster than random positions (AUC gap 0.113 on validation). We also describe an auditable word-level alignment pipeline for raw clips; its outputs do not feed the predictive models.
Abstract (translated)
URL
https://arxiv.org/abs/2610.20543
PDF
-
Neural POS Tagging for the Low-Resource Saraiki
2026-10-14 14:38:17Usman Nawaz, Muhammad Junaid Iqbal, Muhammad Asaf, Mariam Munsif Mir, Faisal Rehman, Tahir Alyas, Muhammad HasnainAbstract
Part-of-speech tagging is a fundamental sequence labeling task and an important component for many downstream natural language processing applications. However, Saraiki, written from right to left in a Perso-Arabic script, remains an under-resourced Indo-Aryan language, despite being spoken as a mother tongue by 28,849,579 people (approximately 28.9 million) according to Pakistan 7th Population and Housing Census 2023. This paper presents a manually annotated Saraiki Universal POS tagging dataset developed from plain-text. We annotated the text with 17 POS categories. We evaluate three pretrained transformer encoders under the same fine-tuning setup: XLM-RoBERTa base, Urdu RoBERTa small, and multilingual DistilBERT. Experimental results over five random seeds show that XLM-RoBERTa base achieves the best performance, with 87.29% +/- 0.21% development accuracy and 88.60% +/- 0.21% test accuracy. Urdu RoBERTa small and multilingual DistilBERT also provide competitive results, suggesting that multilingual and related-script pretrained encoders are effective for Saraiki UPOS tagging. The dataset, evaluation setup, and baseline results introduced in this work provide a benchmark for future research on Saraiki NLP and low-resource sequence labeling.
Abstract (translated)
URL
https://arxiv.org/abs/2610.20279
PDF
-
Choosing the Number of Attention Heads from Intrinsic Dimension: A Small-Scale Study
2026-10-14 10:50:40Arjun Boya, Charan Sai Bandaru, Masthanaiah Pillimetla, Kavana ChidanandappaAbstract
The number of attention heads in a transformer is usually set by convention or grid search. We test a simple heuristic for setting it in advance: estimate the intrinsic dimension of token representations without labels, and use roughly that many heads. We estimate intrinsic dimension with a nested-dropout VAE trained on DistilBERT embeddings of WikiText-2, choosing the KL weight at the knee of the VAE's rate-distortion curve rather than by hand. This gives $k^*=9.2$ (95% CI [8.6, 9.8]). In a small 4-layer transformer trained from scratch, 10 heads and 32 heads are statistically equivalent on SST-2 and on MRPC (paired TOST over 10 seeds, $\pm1$ point margin), with 69% fewer attention parameters for the 10-head model. Training 10 heads from the start also slightly outperforms training 32 heads and pruning to 10 (0.6 points, $p=0.008$). The evidence is narrow: two classification tasks, one small architecture, no pretrained models, and no formal link between intrinsic dimension and head count. We present the result as a practical heuristic worth testing at larger scale, not as a general principle.
Abstract (translated)
URL
https://arxiv.org/abs/2610.19966
PDF
-
Instruction Mining From Images: Constructing a Synthetic Dataset for Multimodal Learning ILSP DCVLR Competition Submission
2026-10-14 10:18:10Eleni E. Kasoura, Dimitrios Damianos, Georgios Paraskevopoulos, Vassilios KatsourosAbstract
This work presents a compact and reproducible pipeline for synthesizing and curating image-based instruction-response pairs for DCVLR. We draw from the Cauldron collection and use Magpie-style instruction and response synthesis with Qwen2.5-VL-7B Instruct to create concise, image-conditioned instruction pairs. The pipeline consists of instruction synthesis, conditioned on input images and controlled via a system prompt, sanitization, tagging, and filtering of the instructions based on the score obtained by a reward model (DeBERTa-v3). We retain samples with positive scores to maximize the quality of the generated instructions. We observe that score distributions have a median and mean reward score of ~ 1.53 and ~ 1.75 respectively, indicating that most generated instructions are of good quality. The resulting dataset contains diverse instructions in the math, spatial, and social reasoning domains.
Abstract (translated)
URL
https://arxiv.org/abs/2610.19932
PDF
-
Physics-Shaped, Not Physical: Auditing Physics-Guided Underwater Image Restoration
2026-10-14 08:17:56Mochamad Rizal Fauzan, Irgi SuryaAbstract
Physics-guided networks for underwater image restoration predict transmission and backscatter and invert the image-formation model, yet rarely test whether this physics does any work. We audit one representative design, PhysGuide-Net, with four controlled experiments: five-seed ablations on UIEB with zero-shot evaluation on the complete LSUI (4279 pairs) and AUIED3K (3000 pairs) sets; a data-efficiency study with a capacity-matched control that removes the formation model; and a validation of the predicted maps against measured range on Sea-thru. With all 800 training images the physics module gives no zero-shot gain (mean PSNR differences of at most 0.16 dB; all 95% intervals include zero) at 3.4x the FLOPs. With 80 images it helps more (+0.65 dB on UIEB, +1.48 dB on AUIED3K; seed-to-seed s.d. on AUIED3K from 1.3 to 0.5 dB), although not individually significant after correction, and the control does not reproduce this (0.8-1.0 dB lower). Yet the predicted transmission does not behave like a physical quantity: its correlation with range is no stronger than for the range map of a different image (pooled difference +0.02, versus -0.06 for the dark-channel prior), and a Beer-Lambert fit explains 0.4% of its variance. The effect of the module also changes sign with degradation severity between datasets. The benefit thus comes from a physics-shaped inductive bias, not from physics. We further show that single-run ablations can reverse sign and document implementation defects that invalidate loss ablations.
Abstract (translated)
URL
https://arxiv.org/abs/2610.19763
PDF
-
NEPSEE: A Double-Annotated Benchmark for Nepali Sentiment Analysis
2026-10-13 20:54:04Gaurab Baral, Mimansak Nepal, Grishma BasnetAbstract
Many sentiment datasets use one annotator for each item, so the test labels show the conventions of that person. We build a Nepali benchmark from two annotators who labelled the same sentences blind. They labelled 5,614 sentences from tweets of Nepali public figures on a five-point scale. We keep only the 3,189 sentences where both give the same polarity. On the three-class benchmark, NepBERTa with a fixed tokenizer gets a macro-F1 of 0.877 and IndicBERTv2 0.867. Three commercial large language models (LLMs) get 0.882 to 0.896 zero-shot, and these differences are not significant. On the five-point scale, 766 sentences have two adjacent labels, for example Pos and SPos. We accept both labels and train on both. On this task, Claude Opus 5.5 beats the best encoders by 0.10 to 0.12 macro-F1. A tokenizer audit adds a second result. The default tokenizers of three Nepali BERT models delete the non-spacing Devanagari vowel signs and the virama. When we restore the marks, macro-F1 increases by 0.084 for NepBERTa but decreases by 0.095 for a model pretrained on stripped text.
Abstract (translated)
URL
https://arxiv.org/abs/2610.19190
PDF
-
Crossing the Gender Binary: Modelling Gendered Reference in Science Fiction and General Fiction
Abstract
Science fiction is widely credited with unsettling gender. This article asks whether that holds at the linguistic level, in how readily gendered reference gives way to its counterpart, a neutral term, or other social category, and whether it separates the genre from general fiction. It does so by masking man/men, woman/women, and person/people across same-period science fiction and general fiction corpora and classifying RoBERTa's five likeliest substitutes by gender. Retention and displacement are permutation-tested over five resampling rounds, and the underlying sentences are read closely. Both corpora cross the gender binary in both directions, but unequally: where woman is masked, the model predicts a masculine term about twice as often as it predicts a feminine term where man is masked. Each mismatch marks a statistical surprise, a word the corpus did not expect there. Against this shared pattern, science fiction shows greater gender permeability than the general fiction corpus, placing man in feminine contexts more often than general fiction while the reverse rate holds steady. Its degendering tendency is stronger too, drawing both gendered terms toward the neutral. Beyond the binary, a masked woman also draws substitutes naming a social category and no gender far more often than a masked man does. In measuring the gender permeability of man, woman, and person in linguistic distribution and testing whether genre changes it, the study gives science fiction studies linguistic evidence for thematic claims and computational literary studies a pipeline applicable to other category vocabularies, such as race and class.
Abstract (translated)
URL
https://arxiv.org/abs/2610.19028
PDF
-
LLM2R: From Large Language Models to 3D Reassemblers
2026-10-13 15:57:41Zhouhan Zhong, Qiuhong Shen, Xingyi Yang, Jinyi Xu, Xinchao WangAbstract
3D part assembly aims to reconstruct a complete shape from a set of unassembled parts by predicting a rigid SE(3) transformation for each part. Because the placement of a part is rarely determined by its own geometry, the task demands reasoning over the entire part set, and existing methods, which rely on pairwise relations or iterative generation and verification, scale poorly to objects with many parts. We present LLM2R, which casts part assembly as sequence modeling with a pretrained decoder-only large language model (LLM): a Point-BERT encoder and a Q-Former compress each part into a block of geometric tokens, the blocks of all parts are concatenated, and a causal LLM adapted with Low-Rank Adaptation (LoRA) processes the sequence in one forward pass. Causal attention, however, lets the hidden states of a part observe only the parts that precede it. LLM2R therefore appends a block of learnable summary tokens after the last part: these tokens follow every part, so their states aggregate the whole assembly at a constant token cost, and a part-aware pose head fuses each part's states with them before predicting its rotation and translation. One architecture thus handles semantic and geometric assembly and predicts all poses in a single non-autoregressive pass. LLM2R achieves the highest part accuracy and success rate on all three PartNet categories, the best rotation error, translation error, and part accuracy on PartNeXt, and the lowest translation error and fastest inference on Breaking Bad, where only a fracture-specialized method is more accurate. Its advantage grows with the number of parts: at 100 parts it exceeds GARF by 33.5 points in part accuracy on Breaking Bad and Assembler by 29.1 points on PartNet, and it transfers zero-shot to unseen categories with more than 16 points higher part accuracy than the strongest of three baselines.
Abstract (translated)
URL
https://arxiv.org/abs/2610.18604
PDF
-
Beyond Black Boxes: Visual Grounding for Text Embeddings
Abstract
In this paper, our goal is to represent text embeddings as sketches for three main reasons. First, to visualize text representation and to explore their potential their potential for AI explainability. Second, to investigate the potential benefits of the sketch-based representation for information retrieval. Third, to explore their potential for reasoning, drawing inspiration from the concept of mental imagery. We propose a sketch-based visual grounding framework that transforms text representations into visual topographic sketches. Our approach aims to emphasis the semantics content encoded in the text embeddings through the corresponding sketch. Evaluations are conducted on the MS-COCO dataset to examine the utility of these sketches for text retrieval. The comparative study of SBERT and Sketch-based method reveals that SBERT retains an advantage in strict top-1 retrieval accuracy while achieving comparable performance to the sketch-based method in top-5 retrieval. A cascaded approach combining SBERT and the sketch-based method outperforms SBERT alone, suggesting that the Sketch-based representation could offer significant potential for information retrieval if trained directly on text.
Abstract (translated)
URL
https://arxiv.org/abs/2610.18047
PDF
-
Taming Softmax at the Limits of Muon: A Unified Approach through Hilbert Geometry
2026-10-13 02:07:50Aditya Somasundaram, Charles Guille-Escuret, Alexander Moreno, Zhengzhong Liu, Eric XingAbstract
Muon controls weight updates through the RMS movement of a linear layer's output. This implicitly treats the linear output as a proxy for the full module, including the nonlinearity that follows. This proxy can be misleading for softmax, where the RMS shift in activations does not faithfully measure the change in the output distribution. We propose Hilbert distance as a natural metric for bounding changes in these output distributions. Hilbert distance bounds changes in cross-entropy loss and limits probability redistribution. For softmax, it takes the form of the variation seminorm of the preactivation perturbation, allowing probability constraints to be expressed as parameter constraints. Formulating the steepest descent problem with a Hilbert distance constraint brings two recently proposed methods under a unified mathematical framework. For the LM head, solving a row-wise approximation to the resulting diameter-constrained problem leads us to RowNorm \citep{pethick2025trainingdeeplearningmodels}. For scaled dot-product attention, the same principle yields the first-order query--key steepest descent problem underlying Compositional Muon \citep{keigwin2026compositional}. These connections motivate using Hilbert distance as a guiding principle for optimizer design in softmax modules.
Abstract (translated)
URL
https://arxiv.org/abs/2610.17588
PDF
-
Assessing the Performance of AI Models in Misinformation Detection Using Predefined Social Noise Constructs
Abstract
Social media platforms generate vast volumes of user-generated content that often obscure meaningful discourse, a phenomenon conceptualized as social noise. While existing misinformation detection methods primarily focus on factual verification, they largely overlook the role of social noise in amplifying misinformation. This study assesses the performance of six text-based models and a graph neural network (GNN) in mitigating the effects of misinformation using predefined social noise constructs. The dataset consists of 500,000 tweets collected from the X platform using the Ukraine hashtag during the Russia-Ukraine conflict from March 2024 to March 2025. Advanced machine learning models achieve strong performance, with accuracies ranging from 79.94% to 97.17%. RoBERTa demonstrates the highest performance, with an accuracy of 97.17%, Cohen's kappa of 0.9660, MCC of 0.9660, and AUC of 0.9977. The GNN achieves an accuracy of 64.52% and an AUC of 0.899, demonstrating performance above random-chance classification by exploiting the relational structure between tweets. All text-based models achieve kappa values greater than 0.75, indicating substantial overall agreement. Explainable AI analysis using LIME reveals that 89.6% of predictive weight is attributed to contextual features rather than isolated keywords. The findings highlight that integrating social noise into AI pipelines enhances misinformation detection and supports more effective mitigation strategies in large-scale social media environments.
Abstract (translated)
URL
https://arxiv.org/abs/2610.17469
PDF
-
Benchmarking Quantum Kernel SVMs for Credit Card Fraud Detection: Performance Realities Under NISQ Constraints
2026-10-12 16:05:42Hitanshi Arora, Mohit Swarnkar, Aswani Kumar CherukuriAbstract
Detecting credit card fraud is notoriously difficult because fraudulent charges make up less than 0.2% of total transactions. Quantum Machine Learning (QML), particularly Quantum Kernel Support Vector Machines (QSVM), promises richer feature representations by embedding classical data into high-dimensional Hilbert spaces. However, whether this theoretical advantage holds up when managing real-world financial data bottlenecks remains unclear. We designed a controlled benchmark comparing three quantum encodings, Angle, ZZFeatureMap, and IQP against classical models while keeping sample sizes, feature counts, and scaling strictly matched. To handle class skew without corrupting quantum state overlaps, we embedded synthetic minority oversampling (SMOTE) directly into the quantum Gram matrix workflow, raising PR-AUC scores by roughly +0.04 across feature maps. When trained on the full European credit card dataset with 30 features, classical baselines reached a top PR-AUC of 0.983. However, restricting classical models to match QSVM hardware constraints (500 samples, 4 PCA features) dropped their PR-AUC to 0.750. This puts our best quantum result 0.730 PR-AUC using IQP encoding on 150 samples into realistic context. Testing on physical quantum hardware reduced performance further to 0.62 due to CNOT gate errors and thermal noise. Overall, our findings pinpoint two main operational bottlenecks for near-term quantum kernels: the scaling of Gram matrix evaluation and heavy information loss from squeezing features onto small qubit registers.
Abstract (translated)
URL
https://arxiv.org/abs/2610.17027
PDF
-
SarcasTraps: Effect of Sarcasm on Turkish Hate Speech Detection
2026-10-12 15:25:27Niyazi Ahmet Metin, Dilara Kek\"ull\"uo\u{g}luAbstract
Standard hate speech detection datasets often emphasize explicit harmful language, but directed harm can also be expressed indirectly, without clearly hateful wording. Sarcasm is one such form of indirect expression, as it can convey harmful intent through mock praise, exaggeration, humor, or apparent politeness. In this paper, we introduce SarcasTraps, a targeted external test set for analyzing the effect of sarcasm on Turkish hate speech detection. SarcasTraps contains 479 entries annotated along two dimensions: hate speech and sarcasm. We train classifiers using BERTurk and XLM-RoBERTa on two existing Turkish hate speech datasets and evaluate them on SarcasTraps, using it strictly as an external test set. We also evaluate GPT-4.1 and Llama-3.3-70B in a zero-shot setting on SarcasTraps. Results show that sarcasm substantially affects hate speech detection. Fine-tuned encoder models lose recall on sarcastic hate, while zero-shot LLMs maintain higher recall but suffer substantial precision drops on sarcastic examples. These findings show that both fine-tuned hate speech detectors and zero-shot LLMs struggle when sarcasm is involved, and SarcasTraps provides quantitative evidence of this challenge through a targeted Turkish test set. Due to the presence of sensitive and hateful language, SarcasTraps will be shared upon request for academic research purposes. Warning: This paper has examples from the dataset which contains offensive and hateful content towards groups of people.
Abstract (translated)
URL
https://arxiv.org/abs/2610.16957
PDF
-
When and How to Canonize: A Generalization Perspective
2026-10-11 13:50:43Yonatan Sverdlov, Benjamin Friedman, Snir Hordan, Nadav DymAbstract
While invariant architectures are standard for processing symmetric data, there is growing interest in achieving invariance by applying group averaging or canonization to non-invariant backbones. However, the theoretical generalization properties of these alternative strategies remain poorly understood. We introduce a theoretical framework to analyze the generalization error of these methods by bounding their covering numbers. We establish a rigorous generalization hierarchy: the error bounds of canonized models are at best equal to the error bounds of structurally invariant and group-averaged models, and at worst equal to the bounds of non-invariant baselines. Furthermore, we show that there exist optimal canonizations which attain the optimal error bounds, and poor canonizations which attain the non-invariant error bounds, and that this depends on the regularity of the canonization. Finally, applying this framework to permutation groups in point cloud processing, we rigorously prove that the covering number of lexicographical sorting grows exponentially with point cloud dimension, whereas Hilbert curve canonization guarantees polynomial growth. This provides the first formal theoretical justification for the empirical success of Hilbert curve serialization in state-of-the-art point cloud architectures. We conclude with experiments that support our theoretical claims. Code is available at this https URL
Abstract (translated)
URL
https://arxiv.org/abs/2605.11008
PDF
-
Toward System One Clinical Intelligence: Parameter-Efficient Fine-Tuning and Counterfactual Explainability of Decision Foundation Models
2026-10-10 16:13:49S. M. Mahbub Uddin AhmedAbstract
Automated disease classification from patient-reported symptoms requires balancing the speed of lightweight discriminative models against the contextual understanding and latency costs of autoregressive large language models (LLMs). Decision foundation models offer an alternative by mapping representations directly to candidate classes without free-form text generation. We adapt Clef-Flash, a 9-billion-parameter decision model built on a Qwen3.5 backbone, for 24-class disease classification on the Symptom2Disease benchmark. Using 4-bit quantized low-rank adaptation (QLoRA) with differential learning rates and a stratified split of 960 training and 240 test cases, the model achieves 95.42% diagnostic accuracy and a macro F1 score of 0.9551. This improves upon the reported zero-shot baselines of 61.67% for a remote API and 65.00% for a local unadapted initialization, and outperforms the ClinicalBERT baseline (90.90%). By evaluating candidate logits in a single forward pass without autoregressive decoding, inference takes under 80 milliseconds per record on a single GPU, yielding a reported 31-fold speedup over generative LLM baselines. To examine symptom-level evidence, we introduce a Leave-One-Symptom-Out (LOSO) counterfactual perturbation framework. Removing disease-defining symptom clauses in a psoriasis example reduces predicted-class confidence from 100.00% to 0.12% and shifts the prediction to impetigo with 99.59% confidence, while removing secondary systemic symptoms produces less than 0.01% change in confidence. These findings suggest that decision-based architectures can combine the linguistic capabilities of large foundation models with low-latency inference and symptom-level interpretability for clinical triage.
Abstract (translated)
URL
https://arxiv.org/abs/2610.15029
PDF
-
UltraBERT: Beyond the Limit of BERT Pre-Training
2026-10-10 16:11:27Jungwoo Park, Taewhoo Lee, Jaewoo KangAbstract
Recent bidirectional encoders have been modernized with innovations from causal large language models (LLMs). However, the pretraining recipe itself remains inherited from small-scale conventions, leaving encoders underoptimized, with no single model performing competitively across both classification and retrieval tasks. In this work, we investigate how far encoder performance can be pushed by revisiting core design choices in training, revealing three key findings: (1) Encoders saturate on downstream tasks beyond a few trillion tokens, indicating that additional compute past this point does not help. (2) Across model sizes, lower masking ratios benefit classification while higher ones benefit retrieval, contradicting prior work suggesting larger ratios for larger models. (3) Distillation with pruning improves classification at low masking ratios but degrades retrieval at high ones. Guided by these findings, we present UltraBERT, a single encoder that assigns separate LM heads to low and high masking ratios within a batch and restricts the distillation objective to the low ratio. To our knowledge, UltraBERT is the first encoder to achieve state of the art on both classification and retrieval tasks, reaching 91.0 on GLUE and 47.7 on BEIR. Together, these results show how far encoder performance can be pushed and provide a principled basis for compute allocation, masking ratio design, and knowledge transfer in bidirectional models.
Abstract (translated)
URL
https://arxiv.org/abs/2610.15028
PDF
