Self-supervised world models are often said to organize their latents around physical quantities. We test this on 2D decaying Navier-Stokes flow in three pre-registered studies, each tested once on held-out data. In Study 1, a joint-embedding predictive architecture (JEPA) trained to predict the flow fails the pre-registered criteria for organizing its latent around fluid quantities better than an untrained network of the same architecture (not supported); a JEPA trained on non-physical linear dynamics and a static invariance model that never sees time do as well or better. In Study 2, plain autoencoders that render the field make enstrophy prominent in their slowest unsupervised latent directions (best r 0.80-0.86 vs 0.38 untrained; fresh seeds, sealed data). Adding VICReg lowers this prominence in every comparison, though not by the pre-registered margin for vorticity, while raising the autoencoders' linear decodability (partially supported). In Study 3, we build steerable directions ("dials") by routing: the decoder gets overall scale only through a small conditioning code, while its other pathway carries a degraded copy of spectral shape. Label-free rules find a loudness dial in all 8 trained models and a shape dial in 3 of 5 fresh seeds on sealed data, against a pre-registered bar of 4; neither control shows a shape dial (partially supported). Across the studies, what a latent contains and what can be controlled through it come apart. In exploration, a supervised direction reading spectral shape almost perfectly moved nothing when steered. During the build, the loudness dial had to be found by its effect on the output (closest to a pure rescaling), not by data variance; for shape, a variance rule and an effect-constrained rule each worked in 3 of 5 seeds. The program used about 70 logged hours on one consumer GPU; all pre-registrations, deviations, code and outputs are public.
Remote sensing change detection is typically formulated as Siamese feature comparison, where two images are encoded independently and a decoder determines which discrepancies correspond to meaningful change. This is challenging because the encoders preserve both true scene changes and nuisance variations such as illumination, seasonal, radiometric, and geometric differences. We introduce ChangeJEPA, a bitemporal joint-embedding predictive architecture that reformulates change detection as latent post-state prediction. Given a pre-change image and partial post-change evidence, ChangeJEPA predicts the full post-change representation under two complementary conditions: a prior hypothesis derived from the pre-change scene and an updated hypothesis conditioned on observed post-change context. A joint-embedding predictive loss aligns both hypotheses with the target latent representation, encouraging semantic consistency while suppressing transient appearance variations. Change is then inferred from the disagreement among the pre-change representation, the observed post-change target, and the predicted latent states, capturing complementary evidence of change, prior surprise, contextual update, and prediction reliability. Across SYSU-CD, LEVIR-CD+, and CDD, ChangeJEPA consistently outperforms its decoder matched Siamese baseline by up to 1.51 IoU points and achieves the best IoU and F1-score on all three benchmarks, demonstrating the advantage of latent post-state prediction over direct feature comparison. Page: https://buddhi19.github.io/ChangeJEPA
Point cloud segmentation is a fundamental task in 3D data analysis, with applications across robotics, augmented and virtual reality, and remote sensing. However, the inherent sparsity, noise, and unordered nature of point clouds introduce significant challenges, particularly in achieving reliable classification. While the state-of-the-art rely on training models for explicit modeling of uncertainty, there is a gap in exploring different roles that estimated uncertainty using generative approaches can play. This work investigates the impact of deterministic uncertainty in point cloud segmentation by explicitly modeling epistemic uncertainty and using it in different roles, namely, as a feature and in loss function. We evaluate a baseline artificial neural network (ANN) and two uncertainty-aware variants (uncertainty as an additional input feature and as a per-sample confidence weight in the training loss) on four large-scale LiDAR benchmarks spanning airborne and mobile acquisition, namely ISPRS Vaihingen, AHN3, Paris-Lille-3D, and Toronto-3D, together with PointNet++ and RandLA-Net references and their uncertainty-augmented counterparts under canonical spatially disjoint train and test splits. Under this protocol, uncertainty-aware modeling improves mean Intersection over Union, recovers minority classes that the baseline ANN never predicts at all, and in half of the configurations improves the alignment between prediction error and estimated uncertainty as measured by Patch Accuracy versus Patch Uncertainty. Our findings indicate that uncertainty can play multiple roles in the segmentation pipeline rather than as a device for improving accuracy alone.
Contrastive vision-language models score left/right relations near chance, but the symptom has two causes that call for opposite repairs: the relation may survive in the pooled embedding yet be unreachable by native image-text similarity, or pooling may discard it while the dense features retain it. On clean two object photographs, across ten encoders, the lateral polarity is linearly decodable from the flip odd component of the pooled vector at 0.77-0.99, yet a search over a million captions leaves nine of the ten at chance under native scoring: an access failure. On controlled C4 orbits, where the relation is the only thing that varies, pooling keeps the arrangement axis at 0.998-1.000 and loses its polarity, which the patch grid retains: a pooling failure. Both sit downstream of the dense features, so one readout covers both. Our zero parameter, query conditioned spatial moment recomputes the geometry from the patch grid and stores it as a vector beside the native embedding. On a 118,287 image index it raises left/right pairwise accuracy against reversed geometry negatives from 0.51-0.53 to 0.81-0.96 and conditional R@1 among 21 category matched candidates from 0.05 to 0.37-0.75, at under a millisecond per batched query, while reproducing the native ranking exactly for queries that name no relation.
Metabolic tumor volume (MTV) summarizes whole-body PET disease burden for prognostic stratification and treatment assessment. Automated MTV is useful only if its estimate is reliable for the individual patient, yet segmentation models are usually evaluated with overlap metrics such as Dice rather than burden error. We examined whether overlap metrics identify unreliable MTV estimates using 1,014 autoPET FDG PET/CT studies. We evaluated LesionTracer predictions held out from fine-tuning, compared five selected released models on 489 tumor-positive patients, and examined the findings in 597 PSMA studies. Across the five FDG models, ranking by higher mean Dice gave the same order as ranking by lower mean relative MTV error ($\rho=-1.00$, exact $p=0.017$). At the patient level, however, out-of-fold Dice tracked relative MTV error strongly ($\rho=-0.75$), yet 82 of 489 patients (16.8\%) fell on opposite sides of the Dice 0.80 and MTV-error 25\% thresholds. Within the training-exposed released models, correlations ranged from $-0.47$ to $-0.74$ over narrower, high-Dice ranges. Out of fold, opposing false-positive and false-negative volumes cancelled about one third of spatial error in net MTV. In 236 of 489 patients, each of the five models missed at least one lesion, though not necessarily the same lesion. Thus, agreement in average model rankings did not establish the reliability of individual MTV estimates or reveal shared lesion misses. PET segmentation benchmarks should report patient-level MTV error and missed-lesion counts alongside Dice and other overlap metrics.
In multi-objective reinforcement learning, preferences may be specified only after data collection, requiring one dataset to support a continuum of reward trade-offs. We study finite discounted Markov decision processes with unknown transitions, $S$ states, $A$ actions, and $m$ known reward functions taking values in $[0,1]$, given access to a generative model. The learner must provide an $\varepsilon$-optimal policy for every convex combination of these rewards without further sampling, with a guarantee that holds simultaneously over all preferences and initial states. Writing $H=(1-\gamma)^{-1}$, we establish the constant-confidence minimax sample complexity $\widetilde\Theta\!\left( \frac{SAH^3}{\varepsilon^2}\min\{S,\log(em)\} \right)$, where the suppressed factors are logarithmic in $S,A,H,\varepsilon^{-1}$ and the dependence on $m$ is explicit. Thus, up to these logarithmic factors, the worst-case statistical cost of supporting all preferences grows logarithmically with the number of objectives and saturates at the number of states. The upper bound uses uniform state--action sampling and empirical planning, with reward rounding in the logarithmic-objective regime. Its analysis chains optimal-value increments over covers of the reward hull. A one-sided Bellman variance bound controls signed value differences under a second transition kernel, preserving the $H^3\varepsilon^{-2}$ dependence. A complementary uniform guarantee for all bounded rewards yields the state-dimension saturation. We establish lower bounds matching up to logarithmic factors through a rare-exit construction with subset rewards and action encoding. These bounds apply to adaptive simulator queries and arbitrary randomized history-dependent output policies, while the upper bounds are attained by deterministic stationary policies.
In biodiversity monitoring, a single image raises several questions: which taxon, how many, which trait, and whether the individual has been seen before. Vision-language models (VLMs) could answer all of these from the image, but their ability is largely untested in biology, as existing benchmarks focus mostly on species identification within a few taxonomic groups. We present Bio-VQA, the largest biological visual question answering (VQA) benchmark and the only one to span four aspects of biological visual understanding: hierarchical taxonomic reasoning, fine-grained attribute grounding, quantitative spatial reasoning, and individual re-identification. Bio-VQA contains 1.86M questions over 217K images from 11 public datasets, and extends right-or-wrong scoring to graded scoring, so a near-miss counts for more than a distant one. Evaluating 18 open-weight models from 3.75B to 38.4B parameters, we find that general visual capabilities do not transfer to biological organisms: the best model counts animals exactly on 41.7% of images with 2 to 10 individuals, less than half its accuracy on everyday objects at the same counts. Taxonomic accuracy falls to one sixth, from 0.72 at phylum to 0.12 at scientific name, and scaling helps mostly at coarse ranks. The gap to a domain specialist widens accordingly, from 0.731 against BioCLIP-2's 0.874 at class (1.2x) to 0.269 against 0.544 at scientific name (2x). On trait size measurement, 15 of the 18 models do no better than an image-blind baseline that predicts one fixed number for every specimen, and on re-identification, 17 of the 18 models fall below chance under adversarial conditions. These failures persist across a tenfold range of model size. We hope Bio-VQA becomes a standard test of how well VLM capabilities transfer to the biological domain.
In this paper, three operations on Gibbs probability measures are studied: renormalization, normalized log-linear combination, and nesting. The resulting measures are shown to be Gibbs probability measures with explicit expressions for their objective functions, reference measures, and regularization factors. These measures are further characterized as unique solutions to optimization problems involving linear combinations of the original objective functions, subject to either relative entropy regularization or relative entropy constraints. When the reference measures are probability measures, geometric interpretations of these three operations are established involving Pythagorean identities and exponential geodesics.A central result is the equivalence between nesting and normalized log-linear combinations of Gibbs probability measures. These results are then applied to supervised machine learning with Gibbs algorithms, which are represented by Gibbs probability measures. In this context, either through nesting or log-linear combinations, Gibbs algorithms trained on local datasets are used to form new Gibbs algorithms trained upon the aggregation of all local training datasets. Interestingly, the parameters of such operations, determine the weights of each training data point on the aggregated dataset. %Operations on local Gibbs probability measures are shown to control the contributions of training data. This observation is leveraged for achieving: (a) exact machine unlearning in centralized, decentralized, and federated learning systems; and (b) centralized-equivalent performance, in the sense of producing algorithms trained on the aggregation of all local datasets in decentralized and federated learning .
In multivariate conformal prediction, many nonconformity scores yield prediction sets with complex geometry (e.g., ellipsoids or density level sets) that are difficult to interpret and visualize. Hyperrectangles offer a practical alternative by providing one interval per output. We extend this notion to Cartesian products of arbitrary one-dimensional sets, which we call generalized hyperrectangles. We introduce a rectangularization framework that turns any continuous multivariate nonconformity score, or conditional density, into one yielding generalized hyperrectangular prediction sets with finite-sample coverage. We establish key geometric properties of our rectangularization framework and provide useful intuitions on generalized hyperrectangles. We further establish a representation theorem showing that a broad class of hyperrectangular conformal methods based on max-aggregated coordinate-wise scores can be interpreted as rectangularizations of implicit conditional densities. Through synthetic and real-world experiments, we quantify the efficiency cost of rectangularization, while highlighting its benefits, especially for practical applications of uncertainty quantification.
LLMs can repair noisy OCR transcripts of historical documents, but they also change text that was already correct. We propose EditVeto, a veto gate placed after a full-context LLM rewrite. The rewrite is character-aligned to the OCR input, its differences are split into blocks, and a 0.6B-parameter gate keeps or reverts each block. An oracle analysis motivates the design: given the true error locations, per-span generation is worse than no correction, whereas vetoing a full rewrite's edits helps. Trained on rewrites by two Qwen3 models and evaluated, frozen, on the eight HIPE-OCRepair-2026 test sets, the gate improves all five rewriters tested, two of them unseen in training: weighted cMER falls by 0.0026-0.0044 and the preference score rises by 0.15-0.25 (paired bootstrap, p ~ 0.0002 each). On published competition outputs it lowers cMER for three of four teams but raises it for the winning run (0.0050 to 0.0065). In a second, disclosed test look, a retrained gate still improves LoRA-tuned rewriters (cMER -0.0011 to -0.0021), and an ensemble of their gated edits reaches a cMER of 0.0069 and a preference score of 0.928, mostly through fine-tuning and pooling rather than gating. Built only on public models and data, it is behind the winner (0.0050) and tied with the second-ranked run (0.0071), both from the only team using domain pre-training, and ahead of all other published runs. A third, secondary look shows the unchanged gate improving zero-shot Qwen3.8-27B (cMER 0.0114 to 0.0108) but only the preference score of its LoRA-tuned version.
Across a wide range of applications, kernel measures are used to tune hyperparameters, select among candidate kernels, and assess the suitability of kernels for specific learning tasks. For classification, kernel measures are used to predict which kernels may produce accurate kernel classifiers. Many existing kernel measures, such as alignment to the ideal kernel, are designed to reward kernels which captures all samples from the same class as similar. While effective in many settings, this requirement is unnecessarily restrictive for classification, where classes may exhibit substantial internal structure. In this work, the "Confidence Score", a new kernel measure, is proposed. Unlike alignment-to-the-ideal kernel measures, the Confidence Score measures the alignment between an encoding of the ground truth and that ground truth projected by the kernel, allowing kernels that preserve intra-class structure to achieve high scores. Through experiments, it is demonstrated that the Confidence Score exhibits a higher correlation with classification accuracy across a variety of kernels on both real-world and synthetic datasets, compared to select other kernel measures.
In this paper, we provide tight lower bounds for minimizing high-order H\"older-smooth and uniformly convex functions when their orders match. Specifically, we consider functions whose $p^{th}$-order derivatives are H\"older continuous with degree $\nu$ and parameter $H$, and that are uniformly convex with degree $q$ and parameter $\sigma$. When these orders match, namely, $q = p + \nu$, we show that any deterministic algorithm with access to derivatives up to order $p$ requires $\Omega\left(\left(\frac{H}{\sigma}\right)^{\frac{2}{3(p+\nu)-2}}\log\frac{\Delta_0}{\epsilon}\right)$ oracle calls to reduce an initial optimality gap of $\Delta_0$ to $\epsilon$. This matches the corresponding upper bound and establishes a higher-order uniformly convex counterpart to Nesterov's classical first-order lower bound for strongly convex functions. Rather than directly generalizing Nesterov's classical quadratic-chain construction, we develop a multiscale phase-wise construction that preserves previous oracle responses while maintaining global H\"older smoothness and uniform convexity. Our result resolves the matched case left open by existing lower bounds for the two asymmetric regimes.
Sparse autoencoders (SAEs) decompose visual representations into individual features, offering a promising basis for explaining model behavior. Interpreting these features typically relies on images or pixel regions that most strongly activate them. However, this evidence becomes ambiguous when multiple features with different semantics frequently activate in the same locations, such as texture and color in the same patterns. We propose Reconstructive SAE Intervention (RecSAEI), a framework that shifts feature explanation from passive observation to active visual reconstruction. By intervening on representations along SAE decoder directions and reconstructing the visual output, we provide direct evidence of a feature's semantic meaning. We demonstrate that our method successfully disambiguates feature meanings across multiple datasets and vision encoders. Furthermore, our framework grounds [CLS] SAE interventions in a sparse set of patch SAE decoder directions. Beyond stable target response, this attribution also exposes the compositional relationship between elementary visual patterns and high-level semantics. Our results advocate for intervention reconstruction as an essential complementary signal for explaining neural networks.
For humans, referential disambiguation is a collaborative process involving clarification questions asked when they will meaningfully change the outcome. However, large language models (LLMs) have been shown to be biased toward premature answering, preferring definite single-turn responses over more clarification exchanges. In this paper, we study the problem of "strategic clarification" -- asking clarification questions in a reference game when their answers will narrow the set of candidate referents. We construct a controlled test environment from the SIMMC 2.1 dataset that records how many questions each ambiguous reference needs to be successfully resolved, and examine how baseline LLM behavior and common post-training methods ask clarifying questions vs. attempt to answer the referent. We also propose BACE (Belief-Aware Critic with Epistemic tracking), a lightweight action chooser method on top of a frozen LM backbone, and show that BACE more consistently chooses the optimal number of clarifying questions than methods such as GRPO and ArCHer.
Textual information is ubiquitous in real-world time series datasets, appearing as news, reports, and annotations that hold key predictive signals not deducible from numerical history alone. Large Language Models (LLMs) interpret text but lack numerical precision, while Time Series Foundation Models (TSFMs) are quantitatively accurate but cannot reason about the effects of text. Moreover, the lack of large-scale paired multimodal datasets limits progress in bridging this gap. In this work, we address this challenge by providing a scalable approach for generating realistic multimodal datasets by annotating real-world time series with LLMs. Building on this, we propose Context as Covariates (CoCo), a forecasting framework in which an LLM distills its textual reasoning into numerically grounded forecast and confidence covariates that guide a TSFM backbone, thereby exploiting the strengths of both LLMs and TSFMs. We align the two models in stages, first fine-tuning the LLM via Group Relative Policy Optimization (GRPO) with a reward tied to covariate-induced improvement in forecasting performance, and then fine-tuning the TSFM on covariates generated by the aligned LLM. Our experiments show that models trained on our synthetic multimodal corpus generalize to real benchmarks spanning finance, economics, weather, traffic, and security. Notably, CoCo with Qwen3-4B-Instruct-2507 as the LLM and Chronos2 as the TSFM achieves up to 12% MAE improvement over state-of-the-art TSFMs and LLM forecasters, such as GPT-4o. Code, checkpoints, and the evaluation datasets are publicly available.
Humanoid motion tracking has scaled to general policies that reproduce diverse human motion, but the resulting behavior is in free-space: the tracker knows where the body should be, not how it should interact with the environment. Other work makes the object part of the objective, yet abstracts that interaction to contact or geometry, which leaves the demanding cases under-explored. We argue that force is a natural command channel between motion imitation and object interaction, one that specifies not just where to make contact but how hard. We present Ctrl-F, a hierarchical motion-force framework that augments motion tracking with an explicit force command, a 3-D force at each designated contact body such as a hand or a foot. The hierarchical framework comprises: (1) A unified motion-force controller tracks the reference motion while applying the commanded force to the world. (2) A high-level force planner that issues a force command online based on object state and proprioception observations. In simulation and on hardware, Ctrl-F completes loco-manipulation that demands large, sustained interaction force and adjusts its command online as the load changes, both out of reach of a force-blind tracker. On hardware, Ctrl-F demonstrates hand-based dragging, pushing, and lifting, as well as foot-based kicking, successfully manipulating boxes weighing up to 20kg in pushing and kicking tasks where the force-blind baseline fails, while adapting online to additional loads introduced mid-task.
Aligning large language models (LLMs) with several objectives, such as helpfulness, harmlessness, and honesty, typically relies on preference judgment for each objective, growing the annotation cost linearly with the number of objectives. Active preference learning can reduce this cost by selecting the comparisons whose labels are expected to change the policy most. Existing remedies either lack an objective-aware query selection mechanism for the multi-objective setting or regard conflicting preferences as harmful, therefore excluding them from the training data. We introduce OGC-DPO (Objective-conditioned Gradient Coverage DPO), an active learning framework that treats each (response pair, objective) combination as a separate query and measures its informativeness through the policy's own implicit gradient, without relying on an external reward model. OGC-DPO then selects the queries that most expand each objective's gradient coverage. Empirical experiments across varying numbers of objectives demonstrate the effectiveness of OGC-DPO over existing active preference learning methods under the same annotation cost. Further analysis demonstrates that informative conflicts, rather than being uniformly harmful, provide a useful policy training signal.
Instantaneous quantum polynomial-time (IQP) Born machines are attractive quantum generative models, as certain IQP circuit families are conjectured to be hard to sample, yet they can be trained classically. Previously proposed conditional IQP Born machines make the parameters of the diagonal layer condition dependent. We introduce conditional IQP Born machines, in which the condition is encoded in the input state and a single IQP unitary is shared across conditions. Efficient sampling and amplitude-ratio evaluation for the reference state entering the diagonal layer suffice to obtain unbiased classical estimates of a conditional maximum mean discrepancy objective and its gradients. We show that input conditioning can represent conditional families that no diagonally conditioned IQP model with the same number of qubits can reproduce. For the sixteen-qubit Curie-Weiss spin model, a single conditional model captures the magnetization transition at held-out temperatures. Its mean total variation distance is six times lower than that of diagonal conditioning under the same protocol and comparable to that of classical generative models with up to seven times as many parameters. We also train conditional models classically with up to 2048 qubits, which reproduce the mean squared magnetization of the target away from the crossover. These results extend classically trained IQP Born machines to conditional generation.
While discrete tokenization in Unified Multimodal Models (UMMs) establishes strong semantic reasoning, it introduces a strict quantization bottleneck that degrades fine-grained visual details, often resulting in structurally accurate but artifact-heavy images. We introduce FlowShift, a parameter-efficient post-training framework that bypasses this bottleneck by converting pre-trained discrete UMMs into continuous rectified-flow generators. Instead of discarding the backbone's semantic priors, FlowShift replaces discrete visual interfaces with continuous latent projections and applies LoRA to the attention layers. To efficiently boost visual quality without semantic drift, we propose Dual-Objective Velocity Vector Distillation (DOVD), which aligns the model with a high-fidelity continuous teacher and ground-truth continuous trajectories. This approach drastically improves perceptual quality while preserving prompt adherence and reasoning. When applied to leading discrete UMMs, FlowShift significantly improves COCO-30K FID and aesthetic scores, while retaining over 99.5% of multimodal understanding capabilities. Ablations confirm DOVD drives generation quality. Ultimately, FlowShift achieves high-fidelity continuous generation without sacrificing pre-trained semantic grounding.
Learning on boundary representations (B-reps) requires robustness to the pose of a part, since CAD models are not stored in a canonical orientation. Existing B-rep networks either operate on world-frame coordinates and normals, and thus degrade under rotation, or encode each face in a local reference frame. Local frames yield invariance, but reduce inter-face messages to scalars and discard the relative orientation of adjacent faces. We propose SO3brep, an SO(3)-equivariant graph neural network on the B-rep face-adjacency graph. Node features are computed in local frames and mapped into a shared frame by Wigner D-matrices. Clebsch-Gordan tensor products couple these features with the directions between adjacent faces, so that relative orientation is preserved throughout message passing and the output remains SE(3)-invariant. Across three face-segmentation and two classification benchmarks with randomly rotated test parts, SO3brep achieves the highest accuracy among the compared B-rep networks and equivariant point cloud networks, with the fewest parameters. Its advantage is largest with few labeled parts, a setting typical of industrial CAD. Code is available at https://github.com/UGent-CVAMO/so3brep.