Peer Revenue Fully funded

Papers

Fresh from arXiv, typeset properly. Highlighted words trade on the market.

Computer Science Machine Learning

cs.LG Watch Machine Learning

Tuesday 21 July

Machine Learning

Contraction-Gauge Preconditioning for Quantized Matrix Multiplication

Piyush Sao, Narasinga Miniskar, Pedro Valero-Lara et al. (+2 more)

We study low-precision computation of C=AB with both factors quantized. We derive an exact finite-dimensional identity for the expected squared product error under independent, zero-mean entrywise errors with known variance fields; it holds exactly for non-overloading subtractive dither and for independent stochastic rounding, and we empirically assess deterministic round-to-nearest (RTN). Using the product-preserving equivalence AB=(AT)(T^{-1}B), we formulate contraction-gauge preconditioning: jointly choosing a factor representation and its sharing pattern before quantization. Preconditioning can reduce product error but may require extra transformed, quantized copies of the opposite operand: a shared transform needs one copy, a block-specific transform up to one per block. Within the bounded family of positive diagonal gauges (folds), a geometric program computes a globally optimal shared fold and a linear program decides whether the identity fold is already optimal. For other families we derive computable selection statistics – tail index for scaling, profile spread for partitioning, coherence and weighted-Gram energy for rotations, slice-energy covariance for hierarchy depth – with upper bounds for ranking heuristic candidates. Across twelve linear products from a trained three-block image classifier, median within-product rank correlations between dither-model predictions and deterministic-RTN errors are 0.937 at 8 bits and 0.918 at 4 bits. The GP fold cuts held-out product error over the identity fold by 18.0% (8-bit) and 20.5% (4-bit) in geometric mean, beats a SmoothQuant-style grid baseline at both precisions and on ten of twelve products, and lowers composed logit MSE by 15.4% and 26.4%. We thus provide exact stochastic product-error accounting, certified selection within the diagonal family, and a common objective for evaluating reusable transform candidates under RTN.
Read the full abstract

All authors Piyush Sao, Narasinga Miniskar, Pedro Valero-Lara, Keita Teranishi, Sudip Seal

Machine Learning

Stale but Stable: Staleness-Adaptive Trust Regions for Stabilizing Asynchronous Reinforcement Learning

Junyao Yang, Yucheng Shi, Zongxia Li et al. (+6 more)

Asynchronous reinforcement learning improves throughput by decoupling rollout generation from optimization, but the resulting staleness is an inevitable byproduct, compounded jointly by policy lag, engine delays, and mixture-of-experts routing. From a trust-region perspective, this mismatch is critical: in the finite-horizon improvement bound, training-inference divergence governs the approximation error, whereas PPO clipping only gates sampled outward updates and therefore acts as a sampled surrogate rather than a full-policy constraint. As a result, the high-staleness update can remain weakly controlled in exactly the asynchronous regime where stale rollouts matter most. We introduce the Staleness-Adaptive Trust Region (SAT), which uses the detached sampled log-ratio as a practical staleness proxy, identifies the high-mismatch tail within each batch through Staleness-based kernel function scaling, and contracts only the sign-selected endpoint of the nominal PPO interval using Effective contraction factors. This design preserves the baseline behavior on ordinary tokens, while making the update more conservative exactly on newly intercepted outward bands. We evaluate SAT in a fully decoupled asynchronous reinforcement learning setup built on Qwen3-30B-A3B-Base, leveraging SGLang as the inference engine and Megatron as the training pipeline. In this setting, SAT-GSPO w/ R3 attains the best observed AIME24 avg@8, reaching 35.83 at lag 1 and 34.79 at lag 8, while SAT-GSPO reaches 34.17 at lag 1. More broadly, the results indicate that aligning the clip interval with observed staleness heterogeneity is an effective way to stabilize the reported asynchronous regime.
Read the full abstract

All authors Junyao Yang, Yucheng Shi, Zongxia Li, Zhongzhi Li, Ruhan Wang, Xiangxin Zhou, Kishan Panaganti, Haitao Mi, Leowei Liang

Machine Learning · Cross-listed from cs.CV

Decoupled Pipeline with Proposal Reranking and Score Fusion for Positive-Unlabeled Marine Species Detection

Robert James Brock, Sebastian Maximilian Krupa and Jason Kahei Tam

The FathomNetCLEF 2026 competition combines underwater object detection and fine-grained marine species classification under a positive-unlabeled evaluation setting. The provided training labels are sparse, while the hidden test set is out-of-distribution relative to the training imagery, creating both annotation incompleteness and source-shift challenges. We describe DS@GT ARC's multi-stage system developed for this setting while keeping model training restricted to the data provided by the competition. The final private-leaderboard model uses a frozen Megalodon YOLOv8x detector as a class-agnostic proposal generator, combines global and tiled inference with tile-edge filtering, classifies expanded proposal crops with a LoRA-finetuned DINOv3 ViT-H classifier, and ranks predictions using weighted geometric fusion of detector and classifier confidence. This system placed 12th out of 102 teams. A closely related variant added a locally trained TTN-inspired validity head as a light reranking signal, improving public-leaderboard and proxy-evaluation performance but slightly reducing private-leaderboard performance. Across experiments, the strongest lesson was that train-derived validation and detector-only metrics were not reliable enough for model selection. Instead, we used proxy datasets only for validation and comparison, and combined those signals with leaderboard feedback and targeted ablations. These experiments showed that reserving proposal recall, avoiding over-aggressive filtering, and improving downstream ranking were more effective than fine-tuning the detector or directly training on noisy pseudo-labels. Code: https://github.com/dsgt-arc/fathomnetclef-2026.
Read the full abstract

Machine Learning · Cross-listed from cs.CV

Attributes Should Come from Images, Not Class Names: Distribution-Conditioned Attribute Selection for Vision-Language Models

Gautam Rajendrakumar Gare, Jia Shi, Zhiqiu Lin et al. (+3 more)

A popular route to interpretable zero-shot classification asks a large language model (LLM) to describe each class name and prompts CLIP with the resulting descriptors. We show that these descriptors carry little visual evidence of their own: removing the class name from the prompt collapses ImageNet accuracy from 59.5% to 15.5%. The diagnosis is that the descriptors are conditioned on the label rather than on the images, so they describe the concept in general and mislead exactly when the data shifts; an LLM insists that strawberries are red, but every strawberry in ImageNet-Sketch is a colorless line drawing. We therefore select attributes from the target image collection instead: we score a large attribute pool against the images in CLIP's joint embedding space and keep the top-scoring attributes per class. Selected this way, class-name-free attribute prompts reach 23.8% on ImageNet (against 15.5% for LLM descriptors), the gain holds on four shifted ImageNet variants, and reselecting from the LLM's own pool isolates the selection mechanism as the cause. With one image per class, the selected attributes outperform the prompt-tuning method CoOp by 3 points while fitting in under a minute instead of 14 hours, with no learned soft prompt to obscure the decision. Because the attribute set is chosen by the data, it doubles as a readable summary of a dataset, which we use to describe distribution shift in words. Our code and results are available on our project page: https://ggare-cmu.github.io/AttributeSelect/
Read the full abstract

All authors Gautam Rajendrakumar Gare, Jia Shi, Zhiqiu Lin, Deepak Pathak, John Galeotti, Deva Ramanan

Machine Learning

Exposure-Based Reinforcement Learning to Rank

Harrie Oosterhuis, Rolf Jagerman, Zhen Qin et al. (+1 more)

Reinforcement learning (RL) methods for learning-to-rank (LTR) can optimize (almost) any ranking goal, e.g., from precision or discounted cumulative gain to fairness-of-exposure or ranking distillation. However, standard RL is ineffective and computationally costly due to the enormous action space in LTR settings. Existing methods reach computational efficiency through custom gradient computation algorithms, but they are very complex to implement and often clash with auto-differentiation. Consequently, existing RL for LTR is not attractive to many practitioners. We reconsider RL for LTR while actively avoiding reliance on custom gradients. Contrary to the existing approaches, we focus on variance reduction and GPU computation. In doing so, we discover that high sample-efficiency can be reached through baseline corrections and partial marginalization. Furthermore, we propose an abstraction that places gradient estimation behind a document-exposure distribution, this enables seamless plug-and-play integration with auto-differentiation. Thereby, one only has to implement a loss as a differentiable function of exposure and RL for LTR can optimize it using auto-differentiation. Our experimental results reveal that our new exposure-based RL for LTR approach converges considerably faster and at significantly higher ranking performance than existing custom gradients, with no additional costs in computation time when using GPUs. In contrast, existing custom gradients result in severe stability issues when converging over many epochs, which never occur for our methods. Thus, we considerably improve RL for LTR methodology by increasing its effectiveness, efficiency, and ease of application.
Read the full abstract

All authors Harrie Oosterhuis, Rolf Jagerman, Zhen Qin, Xuanhui Wang

Machine Learning · Cross-listed from cs.CV

Cross-Dataset Generalization in Breast MRI Tumor Classification via Class-Wise Dataset Mixing

Mohammad Ali Dadrast and Hamid Usefi

Breast MRI is highly sensitive for detecting breast tumors, but exams contain many slices and require substantial reading time. Deep learning models often perform well on internal splits but can fail across institutions because of domain shift and dataset-origin bias. We study this failure mode for binary breast MRI tumor classification. EfficientNet-B3 and WaveViT-Small are trained using Duke Breast Cancer MRI and fastMRI, and evaluated only on the independent multi-center MAMA-MIA cohort. In a deliberately confounded setup, where label is perfectly correlated with dataset origin, external accuracy is near chance (0.5048–0.5265), despite very high recall. We then construct a mixed training set in which each class contains samples from both Duke and fastMRI, while preserving patient-level splitting, augmentation, and leakage controls. On MAMA-MIA, dataset mixing improves accuracy/F1 to 0.8463/0.8625 for WaveViT-Small and 0.8884/0.8994 for EfficientNet-B3. These results show that controlling dataset-origin bias is important for reliable breast MRI classification.
Read the full abstract

Machine Learning · Cross-listed from stat.ML

The Price of Hidden Curvature: Improved Lower Bounds for Bandit Convex Optimization

Nived Rajaraman and Yanjun Han

We establish improved lower bounds on the minimax expected regret of stochastic bandit convex optimization for $1$-Lipschitz functions on the $d$-dimensional Euclidean ball. For time horizons $n\ge d^{10/3}$, we prove a lower bound of $\Omega(d^{4/3}\sqrt{n})$, the first nontrivial bound that exceeds the $d\sqrt{n}$ dependence of linear bandits, showing that stochastic bandit convex optimization is fundamentally harder than linear bandits. For $d^2\le n\le d^{10/3}$, we obtain a lower bound of $\Omega(\sqrt{d}n^{3/4})$, matching the regret of the algorithm of Flaxman et al. (2005), establishing its optimality in this regime. The hard class of convex functions we construct takes the following form in dimension $2d$: for an action $a=(a^1,a^2)\in \mathbb{B}^{2d}$, each function is the scaled soft maximum of a "tube", $r^{-1}\|W^\star a^1-\frac{r}{8\varepsilon}a^2 \|$ (hyperparameterized by $\varepsilon,r$), and a squared distance function, $\frac12\|a^1-u^\star\|^2-\frac12\|u^\star\|^2$. Here $u^\star\in\mathbb{R}^d$ is the unknown target determining the minimizer, while $W^\star\in\mathbb{R}^{d\times d}$ hides the region in which the quadratic curvature is observable. Indeed, observations reveal substantial information about $u^\star$ only when the learner acts near the hidden tube $a^2\approx \frac{8\varepsilon}{r}W^\star a^1$; away from it, the tube branch masks the quadratic branch. Thus the learner must pay to uncover the geometry encoded by $W^\star$ before it can effectively exploit the curvature that identifies $u^\star$. Formalizing this tradeoff yields a sample complexity lower bound of $\Omega(\frac{d^{5/2}}{\varepsilon^2}\wedge\frac{d^2}{\varepsilon^4})$ for finding an $\varepsilon$-optimal action, and ultimately the $\Omega(d^{4/3}\sqrt{n}\wedge\sqrt{d}n^{3/4})$ regret lower bound. The proof was developed by GPT-5.5 Pro and GPT-5.6 Sol Pro under the authors' guidance.
Read the full abstract

Machine Learning

Spaghetti Architect: A Contamination-Resistant, By-Construction-Labelled, Multi-Language Code Dataset Generator

Yuxiang Ji

Mined code corpora are abundant but uncontrolled: a snippet's semantics, surface "messiness," and difficulty are whatever the wild contained; there is no known-optimal reference to grade against; and any public sample may already sit in a model's training set. We present Spaghetti Architect, a tool that mints code datasets with the control such corpora lack. An anti-optimization transpiler maps a clean, language-agnostic JSON intermediate representation to deliberately redundant, fully-flattened programs in five languages (Python, JavaScript, Go, Java, C++); every program is compiled, run, and checked against a reference oracle, so each instance is correct by construction. The clean IR is a known-optimal reference, messiness is dialed by strictly-nested anti-pattern profiles, each instance is labelled along two orthogonal difficulty axes, intrinsic (problem size) and incidental (presentation at fixed semantics), and contamination is resisted by minting fresh variants from a private held-out seed. We give construct-validity evidence that the quality order moves established complexity and readability metrics, and report baselines on a four-model open ladder: exact match rises with scale, and the intrinsic knob collapses arithmetic-aggregation accuracy of even the strongest model to zero. Further, development-set scores equal freshly re-minted held-out counterparts within $|\Delta|\le 0.012$ (comprehension) and $\le 0.011$ (refactoring); on identical programs, refactoring equivalence ($0.73 \rightarrow 0.99$) is scale-invariant while output prediction collapses; and ablating the generator's self-annotations shows they inflate the weakest model an order of magnitude more than the strongest ($-0.173$ vs $-0.017$): the annotated ladder resolves one of three adjacent pairs where the unannotated resolves all three. Open source (MIT), dependency-free, archived under a persistent DOI.
Read the full abstract

Machine Learning

Mark, Don't Erase: Token Inoculation for Dual-Use Knowledge in LLMs

Seunghyun Lee, Dongyoon Han and Sangdoo Yun

In a dual-use field, hazardous knowledge and the benign expertise around it share vocabulary, concepts, and causal structure. That overlap makes selective refusal hard, and existing methods such as unlearning and refusal training overgeneralize their suppression into the benign side. We argue for conditioning, not suppression. In this work, we introduce Token Inoculation, which marks hazardous sentences with a special token, the inoculation token, during continued pre-training. We then fine-tune the model to refuse hazardous questions and, on the same questions under the inoculation token, to predict the hazardous answers. At test time the inoculation token is absent, so the model does not express the hazardous answers it learned to predict. Across 1.5B to 14B parameters and three architecture families it holds hazardous accuracy near 18% while retaining 93% of the instruct baseline's MMLU performance, the best safety-utility balance among the baselines we compare. The same advantage reproduces under open-ended generation and in a cybersecurity transfer. We test this with a causal do(X) analysis that randomizes the accuracy of the conditioning signal, and find that refusal becomes more selective the more of the hazardous domain the model still knows. Our findings underscore the importance of training to predict the responses that must be refused, and outline a direction for safety alignment that does not degrade models' growing frontier knowledge. Project page: https://token-inoculation.github.io/
Read the full abstract

Machine Learning · Cross-listed from cs.RO

End-to-end Conditional Diffusion for Realistic and Controllable Visual Traffic Scenario Generation

Jingzheng Li, Yufei Ge, Zhijun Chen et al. (+8 more)

Generating closed-loop traffic scenarios that are both realistic and controllable is crucial for evaluating autonomous driving systems, especially under rare safety-critical interactions. Existing learning-based methods often struggle to balance controllability and realism, offering either limited fine-grained control over traffic behavior or controllable scenarios at the expense of behavioral plausibility. This paper presents E2E-CDiff, an end-to-end conditional diffusion framework for controllable and realistic scenario generation. Conditioned on front-view visual observations, E2E-CDiff jointly denoises future motion states and executable low-level controls for route-interacting background vehicles. This unified state-action generation mitigates the planning-control mismatch in conventional two-stage trajectory-then-controller pipelines. Differentiable guidance further regulates speed, enforces drivable-area compliance, and supports collision-avoidance or collision-seeking behaviors, enabling both naturalistic and safety-critical scenario generation. Experiments on Bench2Drive show that E2E-CDiff achieves a favorable controllability-realism trade-off compared with representative reinforcement- and imitation-learning baselines, while its collision-guided variant induces challenging interactions across multiple autonomous driving systems. E2E-CDiff also performs competitively as a learning-based ego planner, demonstrating the generality of end-to-end state-action diffusion.
Read the full abstract

All authors Jingzheng Li, Yufei Ge, Zhijun Chen, Qianren Mao, Zizhe Wang, Binhang Qi, Bing Li, Keyu Chen, Baochang Zhang, Xianglong Liu, Philip S Yu

Machine Learning

Graph Neural Network-based Algorithm Selection for the Traveling Salesman Problem: A Systematic Study of Cost and Rank Losses under Distinct Budget Regimes

Zhaoxuan Li, Jiale Yang, Yifei Lu et al. (+1 more)

Automated Algorithm Selection (AS) aims to improve problem-solving performance by selecting, for each problem instance, the most suitable algorithm from a predefined portfolio. This is particularly relevant to the Traveling Salesman Problem (TSP), where solver performance is strongly instance-dependent. We introduce GNNAS-TSP, a Graph Neural Network (GNN)-based AS framework that learns TSP instance representations directly from raw graph data, avoiding manual feature engineering. GNNAS-TSP formulates AS as a joint cost-prediction and ranking task. We evaluate cost-based (mean squared error (MSE), mean absolute error (MAE), and Huber), rank-based (RankNet, ListNet, and LambdaRank), and hybrid learning objectives for a portfolio comprising Chained Lin-Kernighan, Edge Assembly Crossover, Lin-Kernighan-Helsgaun, Multiagent Optimization System, and Concorde. Experiments use fixed computational budgets of 10 and 60 seconds. On the held-out test set, the selected configurations improve on the Single Best Solver (SBS) in normalized solution cost at both budgets. For the 10s budget, AS achieves substantial and statistically significant cost improvement over SBS. Overall, the results suggest that GNNAS-TSP is a useful meta-solving strategy when exploitable variation exists across solver performance.
Read the full abstract

All authors Zhaoxuan Li, Jiale Yang, Yifei Lu, Mustafa Misir

Machine Learning · Cross-listed from cs.CL

LatentMT: Machine Translation with Latent Reasoning

Wei-Rui Chen, Samar M. Magdy, Chiyu Zhang et al. (+3 more)

Latent-reasoning looped language models (LoopLMs) offer a different scaling path for machine translation (MT): instead of increasing parameter count or emitting explicit chain-of-thought tokens, they spend additional recurrent computation inside hidden states. We introduce LatentMT, the first systematic study of latent-reasoning LoopLMs for machine translation. LatentMT adapts a small 2.6B-parameter backbone model with lightweight training. Across 32 translation directions spanning high-, mid-, and low-resource languages, LatentMT achieves performance comparable to models three to five times larger. It is competitive in a high-resource language and achieves state-of-the-art performance on both mid-resource and low-resource languages. Studying the behavior of scaling the number of recurrent reasoning steps, we find that recurrent computation consistently improves translation quality in early steps, then saturates quickly afterwards. Our mechanistic analysis shows that hidden-representation differences shrink along the recurrent reasoning-step axis, supporting the observed saturation in performance. Finally, our efficiency analysis shows that LatentMT requires lower training and inference compute than much larger non-latent-reasoning models with similar performance, making latent recurrent computation a promising path toward compact, efficient, and strong machine translation.
Read the full abstract

All authors Wei-Rui Chen, Samar M. Magdy, Chiyu Zhang, Wenhui Zhu, Zhipeng Wang, Muhammad Abdul-Mageed

Machine Learning · Cross-listed from cs.CL

Stochastic Meta-Unlearning: Bridging Language Backbone and Multimodal Unlearning

Zijie Liu, Jinhao Duan, Gaowen Liu et al. (+2 more)

Machine unlearning for vision-language models (VLMs) remains underexplored. Unlike language models, VLMs combine a language backbone with visual components, which makes unlearning more complex. There is a surprising phenomenon when moving from single-modality unlearning to VLM unlearning: a target forgotten by the standalone language backbone can still be recovered when image information is given to the full VLM. This shows that text-only feedback is not enough for reliable VLM unlearning. Motivated by this observation, we propose Stochastic Meta-Unlearning (SMU), a bilevel framework that uses VLM-level feedback to learn an unlearning-ready initialization. In the inner loop, SMU applies a few unlearning steps to the language backbone using text data. In the outer loop, SMU recomposes the updated backbone with the frozen VLM and evaluates forgetting and utility at the VLM level. This design makes the unlearning update aware of the final multimodal behavior, while still keeping the update local to the language backbone. Experiments on two VLMs, two multimodal meme datasets, and three baselines show that SMU achieves the best overall forget-retain trade-off. Compared with the strongest baseline for each metric, SMU reduces average Forget accuracy by 10.52 points and improves average Retain and Test accuracy by 20.10 and 17.01 points, respectively. More importantly, SMU also transfers to new forgetting targets and to different meta-test unlearning methods. These results suggest that VLM-level feedback can make language-backbone unlearning more reliable and more transferable for VLMs.
Read the full abstract

All authors Zijie Liu, Jinhao Duan, Gaowen Liu, Sijia Liu, Tianlong Chen

Machine Learning · Cross-listed from cs.RO

Intelligent Multi-UAV Navigation in ITNTNs: A Hierarchical LLM Approach

Zijiang Yan, Hao Zhou, Wael Jaafar et al. (+4 more)

The deployment of high-speed Uncrewed Aerial Vehicles (UAVs) in 3D aerial highways necessitates robust coordination of physical flight kinematics and multi-tier network handovers. While Deep Reinforcement Learning (DRL) offers rapid tactical control, it lacks the zero-shot strategic reasoning required to quickly adapt to dynamic Integrated Terrestrial and Non-Terrestrial Networks (ITNTNs). Conversely, Large Language Models (LLMs) excel at semantic reasoning but suffer from high inference latency, rendering them unsuitable for real-time aerodynamic control. To bridge this gap, we propose a novel Hierarchical LLM-driven control framework. A massive cloud-based LLM deployed on a High-Altitude Platform Station (HAPS) manages slow-timescale global load balancing, while lightweight edge-LLMs on individual UAVs translate local observations into tactical sub-goals. These sub-goals guide a fast-timescale physical DRL controller to execute collision-free, handover-aware trajectories. Simulation results demonstrate that our agentic architecture significantly reduces collision rates and improves aggregate system throughput compared to existing baselines.
Read the full abstract

All authors Zijiang Yan, Hao Zhou, Wael Jaafar, Jianhua Pei, Ping Wang, Halim Yanikomeroglu, Hina Tabassum

Machine Learning

BRIDGE: Bottleneck-Aware Regulator-Set Inference and Diagnosis for Cooperative Gene Regulatory Recovery

Maryam Rahimimovassagh, Clayton Thomas Barham, Ivan Garibay et al. (+1 more)

Cooperative gene regulation often depends on groups of regulators acting jointly, but most gene regulatory network (GRN) inference methods output pairwise regulator-target rankings. We introduce Bottleneck-Aware Regulator-Set Inference and Diagnosis (BRIDGE), a framework for complete regulator-set recovery, and Targeted Recovery Attribution for Cooperative Evaluation (TRACE), a diagnostic suite that attributes failures to retrieval, set-level scoring, decoding, and evaluation bottlenecks. TRACE includes a leak-free mechanism-mismatch cooperativity stress test in which cooperative targets are generated by random nonlinear mechanisms rather than product interactions. This design avoids feature-mechanism circularity: Residual higher-order set scoring (Residual HOS2) operates on raw expression vectors without handcrafted product-correlation features. Across 30 matched seed-cooperativity settings, Residual HOS2 improves Jaccard similarity from 0.382 to 0.460, recall from 0.522 to 0.597, and exact recovery from 0.053 to 0.113 over a decomposable pairwise set scorer (PairS2), although exact recovery remains low. On SERGIO DS3, oracle retrieval and TRACE show that candidate coverage is necessary but insufficient because set-level misranking remains the dominant source of exact-recovery failure. PairS2 proposal followed by Residual HOS2 reranking reduces HOS2-scored candidate sets by 94-97% while largely preserving exact-recovery behavior. These results distinguish edge ranking, candidate retrieval, set-level scoring, and exact cooperative regulator-set recovery as separate objectives.
Read the full abstract

All authors Maryam Rahimimovassagh, Clayton Thomas Barham, Ivan Garibay, Niloofar Yousefi

Machine Learning

A Self-Evolving Default Action for Cooperative Tasks with Continuous Action Space

Shuangyao Huang

Counterfactual credit assignment has proven effective in multi-agent reinforcement learning (MARL) for discrete action spaces, yet its extension to continuous-action cooperative tasks remains challenging. Existing methods that approximate the counterfactual baseline via Monte Carlo sampling often introduce bias into policy gradients and fail to guarantee convergence to local optima, as the sampled actions may not have been sufficiently trained. To address these limitations, we propose SAFE, a novel MARL framework that employs a counterfactual baseline conditioned on a self-evolving default action sampled from each agent's experience buffer. This design naturally extends to continuous action spaces without relying on additional simulations, reward models, or environment-specific prior knowledge. The baseline accurately quantifies each agent's contribution, and introduces no bias into the deterministic policy gradient, ensuring convergence to local optima. Extensive experiments on cooperative vehicular tasks demonstrate that SAFE consistently outperforms state-of-the-art models.
Read the full abstract

Monday 20 July

Machine Learning · Cross-listed from cs.CL

Sampling Reveals Style: Unsupervised, Training-Free Discovery of Prompt-Conditional Stylistic Axes in LLM Activations

Ajit Mallavarapu and Ziwei Gu

Large language models (LLMs) encode rich stylistic structure in their hidden activations, but discovering which stylistic dimensions are salient for a given prompt typically requires supervised contrastive data. We present a training-free, prompt-conditional alternative: we repeatedly sample completions of a single prompt at elevated temperature, apply Principal Component Analysis (PCA) to the pooled hidden activations, and label the resulting axes automatically from the pole generations. We validate the discovered axes against 245 human-elicited stylistic annotations in a two-phase study. On our strongest model (Qwen-3.5-4B-Instruct), the top two axes match spontaneously requested human dimensions with 72.8% precision and 43.6% macro-recall, and 75.6% of validity ratings judge the axes' polar generations accurate to their labels, with 90.9% adjacent inter-annotator agreement. Discoverability is strongly model-dependent: both Qwen models and Llama-3.2-3B expose human-salient axes, while DeepSeek-7B-Chat drops to 35.3% precision, its leading components dominated by structural rather than stylistic variance. Simple PCA over a model's own decoding variance is thus an effective, low-cost probe of stylistic structure in LLM representations, one that also exposes sharp cross-model differences in how that structure is organized.
Read the full abstract

Machine Learning · Cross-listed from cs.CL

Subliminal Prompting Beyond Static Geometry: Causal Depth and Multi-Token Confounds

Barath Velmurugan

Subliminal learning shows that language models can transmit a hidden trait through outputs that appear unrelated to it. One proposed explanation, token entanglement, links animal and number tokens through the model's output vocabulary. Yet existing measurements answer different questions: whether outputs co-vary, fixed output vectors align, an answer can be read from a hidden state, or that state causally controls the answer. We measure each separately in a fixed animal-number prompting protocol. From Llama-3.1-8B to 70B, fixed output-vector similarity predicts behavior less well: the paired mean correlation change is -0.080 (95% CI [-0.127, -0.035]). A fixed output-head readout shows no resolved change in normalized depth AUC. To test control, we copy the temporary answer-position state from one number prompt into another at five depths and measure which prompt the final animal score follows. Donor-control AUC rises from 0.254 to 0.540, a paired change of +0.286 (95% CI [+0.272, +0.300]), with increases for all 18 concepts. The contrast remains with exactly eight transformer blocks remaining, while specificity and identity controls remain small or exact. In two Qwen models, scoring every digit in sequence does not recover the positive one-token association. Per-token averaging instead creates a positive pooled association that disappears after controlling number width, revealing a length confound. Thus, fixed geometry, observational readability, causal timing, and multi-token measurement are distinct properties of this frozen prompting channel. They constrain token-level explanations but do not identify the mechanism of training-time trait transfer.
Read the full abstract

Machine Learning · Cross-listed from physics.comp-ph

LSR-Net: Learning the Forward Evolution Operator for Nonlinear Fluid Dynamics

Qian Hou, Sutrisno, Yuqing Li et al. (+1 more)

We introduce the Long-Short-Range Neural Network (LSR-Net), a novel neural operator architecture designed for data-driven forward evolution modeling, and extends it to the prediction of nonlinear fluid dynamics. LSR-Net learns the evolution operator of a dynamical system solely from pairs of initial and future state snapshots, which splits the learnable integral kernel into long-range (LR) and short-range (SR) components within stacked network blocks. While the SR component uses standard convolutions to capture local dynamics, the LR component employs a sum-of-exponentials (SOE) representation. This allows for the efficient computation of global interactions as a trainable Fourier multiplier, reducing computational complexity to $O(n \log n)$ where $n$ is the number of pixels in an input snapshot and requiring only a few parameters per channel. LSR-Net is evaluated on three challenging 2D benchmarks: the coupled Burgers equation, the wave equation with a spatially varying coefficient, and the nonlinear shallow water equation {(SWE)}. Results demonstrate that LSR-Net significantly outperforms the baseline short-range network (SR-Net) as well as FNO and DeepONets in predictive accuracy, achieving substantially lower relative errors by effectively capturing both local fine-scale structures and crucial global pattern interactions.
Read the full abstract

All authors Qian Hou, Sutrisno, Yuqing Li, Zecheng Gan

Machine Learning

Tabular Deep Learning vs Classical Machine Learning for Urban Land Cover Classification

Muntasir Tabasum, Tanpia Tasnim, Md. Ekramul Islam et al. (+1 more)

Urban Land Cover (ULC) classification plays a crucial role in urban planning, environmental monitoring, and sustainable development. We study this task using the ULC dataset from the UCI Machine Learning Repository, which includes tabular features derived from high-resolution aerial imagery across nine classes (e.g., roads, trees, grass, water). The dataset presents typical remote sensing challenges, including high dimensionality, heterogeneous features, and class imbalance. In a unified, reproducible pipeline, we benchmark classical machine learning models (e.g., Logistic Regression, SVM, Random Forest, XGBoost, CatBoost) against Tabular Deep Learning (TDL) models (TabNet, FT-Transformer, TabTransformer, TabSeq, and 1D CNNs). To address class imbalance, we employ weighted cross-entropy loss for TDL models and evaluate performance using accuracy, macro-precision, macro-recall, macro-F1, AUC-ROC, and confusion matrices. Our results show that while tree ensembles remain strong general baselines, TDL models can match or exceed their performance when non-linear interactions are significant and imbalance handling is effective, providing complementary advantages for urban land cover mapping. See code: https://github.com/mtesha/tdl-vs-ml-urbanlandcover
Read the full abstract

All authors Muntasir Tabasum, Tanpia Tasnim, Md. Ekramul Islam, Al Zadid Sultan Bin Habib

Machine Learning · Cross-listed from cs.AR

WARD: Runtime Workload-Adaptive Vision TRansformer Framework for Dependable Edge AI

Mahdi Taheri, Pramit Kumar Bhaduri, Mohammad Masoumi et al. (+1 more)

Edge-deployed AI operate under dynamically changing power budgets, reliability requirements, and input distributions, requiring continuous adaptation. Such conditions arise in long-running edge AI applications, including autonomous systems, industrial monitoring, and satellite onboard intelligence. Existing fault-tolerant methods assume static operating conditions, whereas continual learning techniques neglect concurrent hardware faults during online adaptation. Moreover, the practical deployment of runtime-adaptive reliability frameworks on programmable AI accelerators remains largely unexplored. This paper presents WARD, a runtime-adaptive Vision Transformer framework that combines channel-wise subnetwork partitioning, reliability-aware continual learning, and dynamic operating-mode scheduling to jointly optimize performance, fault tolerance, and adaptation according to runtime conditions. Two physically isolated subnetworks execute under four operating modes (i.e. Full-Precision Mode, Low-Power Mode, High-Reliability Mode, and Adaptive Mode) that dynamically adjust computational cost and reliability while ensuring uninterrupted inference for real-time requirements. To validate the practical deployability of the proposed framework, WARD is implemented on a lightweight FPGA-based accelerator extended with runtime hardware support for mode scheduling and resource management. Experimental results demonstrate that the proposed split architecture achieves a network-level failure rate of only 1.79% under high Bit Error Rates. The hardware implementation incurs less than 5% area overhead and supports runtime mode transitions within few clock cycles, demonstrating that adaptive reliability management can be integrated into programmable edge AI accelerators with negligible implementation overhead.
Read the full abstract

All authors Mahdi Taheri, Pramit Kumar Bhaduri, Mohammad Masoumi, Ali Mahani

Machine Learning · Cross-listed from cs.AR

REQAP: Resilient Weight Packing and Quantization for Edge DNN Acceleration

Mahdi Taheri, Samira Nazari, Mubassher Ansari et al. (+4 more)

Efficient deployment of Deep Neural Networks (DNNs) on edge accelerators requires aggressive model compression while maintaining reliability in fault-prone hardware environments. This paper presents a reliability-aware quantized weight packing methodology for systolic-array-based DNN accelerators. A sensitivity-driven mixed-precision quantization framework assigns layer-wise bit-widths according to accuracy impact while enforcing symmetric precision between weights and activations. A deterministic register-level packing strategy consolidates multiple heterogeneous operand pairs into fixed-width register words, enabling SIMD-within-a-register (SWAR) style parallel execution that reduces both memory footprint and execution cycles. To improve resilience against hardware faults, selective bit-level protection replicates the most significant bits (MSBs) of critical layers into unused register space, achieving TMR-style protection with minimal overhead. A systolic-array simulation framework is developed to evaluate the proposed packing and fault-tolerance mechanisms under realistic execution conditions. Simulations in AlexNet, VGG-11, and ResNet-18 demonstrate up to 62% memory reduction and up to 56% reduction in Multiply-Accumulate (MAC) operations, while significantly improving accuracy resilience under fault injection compared to baseline and fully protected models.
Read the full abstract

All authors Mahdi Taheri, Samira Nazari, Mubassher Ansari, Ali Azarpeyvand, Mohsen Afsharchi, Maksim Jenihhin, Christian Herglotz

Machine Learning · Cross-listed from cs.IR

RecEvolve: A Knowledge-Driven Autonomous Agent System for Recommender Systems

Weidi Pan, He Ma, Shuhao Ye et al. (+6 more)

The rise of agentic AI has catalyzed a shift toward self-iterating systems, opening new frontiers for the autonomous optimization of production recommender models. This paper presents the empirical validation of a knowledge-driven autonomous agent system, deployed directly on a production large-scale Two-Tower retrieval model. By delegating the entire research lifecycle, spanning idea generation, code implementation, offline training, and metric evaluation, to a continuous closed-loop autonomous framework, the agent system executed over 40 completed autonomous training runs from scratch. Executing these runs under rigorous production-scale evaluations, the system systematically navigated hidden architectural bottlenecks on the latest production model to achieve a breakthrough ~20% relative improvement in NDCG, a gain that translated directly to a +3.77% increase in user satisfaction in live production traffic. Furthermore, the deployment exposed critical vulnerabilities in standard evaluation protocols, as the agent system autonomously discovered reward-hacking shortcuts. These findings prove that an autonomous pipeline can dramatically accelerate the pace of machine learning research and stress-test the rigorousness of underlying experimental infrastructure, while also exposing novel challenges such as reward hacking and redundant exploration of failed hypotheses.
Read the full abstract

All authors Weidi Pan, He Ma, Shuhao Ye, Palaksh Rungta, David McPeek, Junyi Jiao, Arnab Bhadury, Mingyan Gao, Onkar Dalal

Machine Learning

Model of Models: When Does Emitting a Specialist Beat Attending, Adapting, or Tuning?

John C. Howell

Given a task described by a few examples, how should a model be specialized to it? Four mechanisms are available – zero-shot, in-context attention, test-time gradient adaptation, and emitting specialist weights from a hypernetwork – yet the operating regime of the last is rarely mapped. We run the identical four-way comparison across six tasks spanning regression, generation, language modeling, reinforcement learning, and clinical and genomic classification, holding the specialist, the context, and (where we can) the training budget fixed. The clearest wins for emission are about cost at matched quality: it ties the state-of-the-art amortized tabular model (TabPFN) on clinical few-shot classification while emitting a reusable specialist instead of re-attending the support set per query, and reaches noise-floor shape generation with a $132$-float per-instance program. On few-shot sinusoid regression it is $2$–$3$ orders of magnitude below MAML at zero test-time gradient steps – a margin that narrows to $\sim$$30\times$ but persists once training budgets are equalized. Emission cannot match in-context attention on high-dimensional sequence modeling: under matched-budget pre-training a one-pass adapter recovers only a minority of the in-context gain ($14.0\pm0.9\%$ at $5$M, $11.2\pm0.5\%$ at $15$M), and a LoRA-rank sweep shows this shortfall is a partial capacity limit – capture climbs from $5\%$ to $21\%$ as rank grows but plateaus far below full recovery. Mechanism ablations confirm the emitted specialist is genuinely task-conditioned, not a memorized prior; and, more speculatively, emitted specialists compose in weight space – interpolating two of them tracks the corresponding blend of their functions. We close with a falsifiable thesis, operationalized through a per-task resolution measure, bounding when each conditioning mechanism should be preferred.
Read the full abstract

Machine Learning · Cross-listed from cs.CL

LLM Safety Alignment in Low-Resource Languages: A Systematic Literature Review

Valdini Douglace Lemofouet, Blessing Ngozi Uzor, Paula Chikaodinaka Anyanwu et al. (+9 more)

Large Language Models (LLMs) have achieved substantial progress in safety alignment, yet their safety guarantees remain significantly weaker in low-resource and multilingual settings than in high-resource languages. In this paper, we conduct a Systematic Literature Review (SLR) of LLM safety alignment in low-resource languages by adopting the PRISMA 2020 methodology. Out of roughly 1,500 papers identified from Semantic Scholar, arXiv, and OpenAlex, 50 relevant studies have been selected and analyzed. Our review is organized around four themes: safety alignment methods, multilingual safety risks, evaluation benchmarks, and cross-lingual transferability. We further propose a taxonomy of safety alignment approaches based on three adaptation mechanisms: data adaptation, objective optimization, and mechanistic alignment. Across literature, translated English benchmarks fail to sufficiently represent culturally rooted harms, and multilingual models are more vulnerable to cross-lingual jailbreaks, code-switching attacks, and safety degradation in underrepresented languages. These failures are driven by several key factors, including uneven multilingual pre-training coverage, insufficient native-language preference data, poor transfer of safety representations, and a lack of culturally aware evaluation frameworks. The review also notes that many low-resource languages, especially African languages, have fewer safety benchmarks available than other multilingual regions. Overall, the results reveal a persistent multilingual safety gap, and suggest that future progress will require culturally grounded benchmarks, participatory data collection, balanced multilingual pre-training, and scalable multilingual alignment methods.
Read the full abstract

All authors Valdini Douglace Lemofouet, Blessing Ngozi Uzor, Paula Chikaodinaka Anyanwu, Danielle Blanche Kapsa, Sukairaj Hafiz Imam, P Sam Sahil, Abigail Oppong, Tassallah Abdullahi, Clemencia Siro, Idris Abdulmumin, Seid Muhie Yimam, Shamsuddeen Hassan Muhammad

Machine Learning · Cross-listed from cs.CL

Evaluating Federated Pre-Training: On the Reliability of Downstream Fine-Tuning and Intrinsic Evaluation

Claudia Grosser, Maike Heuer, Denis Krompass et al. (+1 more)

Federated pre-training offers a way to train foundation models on private or distributed data without centralizing the underlying datasets. However, evaluating federated pre-training remains challenging because differences in client participation and local data availability can make directly comparable evaluation difficult. Moreover, pre-training test perplexity is tied to the pre-training distribution, while downstream benchmarks introduce task-specific adaptation that may not faithfully reflect the test perplexity established during pre-training. In this work, we study which evaluation protocol more reliably reflects federated pre-training quality. Using a controlled set of centralized and federated-trained models of a 16M parameter transformer model trained on identical client data, we assess evaluation protocols by whether they preserve a reference ranking established on the same pre-training testset. We compare downstream fine-tuning on GLUE, including full, head-only, and reduced-data variants, with next-token prediction on GLUE text as an intrinsic evaluation signal. Our results show that downstream fine-tuning does not reliably preserve the pre-training ranking, whereas direct next-token prediction exhibits a strong correspondence with the pre-training test perplexity. These findings suggest that downstream fine-tuning alone can be misleading when comparing federated pre-trained models, and that evaluation signals closer to the original pre-training objective deserve greater attention.
Read the full abstract

All authors Claudia Grosser, Maike Heuer, Denis Krompass, Thomas A. Runkler

Machine Learning

CORVUS: Context Optimization and Reduction Via Underlying Synchronization for LLM Coding Agents

Mingwei Zheng, David OBrien, Siwei Cui et al. (+4 more)

LLM coding agents operate by constructing trajectories that accumulate reasoning, tool calls, and results to enable multi-step decision-making. However, the conventional append-only trajectory architecture found in practice tightly couples file-read actions with their observations, capturing snapshots that become permanently fixed in the chronological history. As files change through agent edits or concurrent human modifications, these snapshots become stale, causing reasoning errors and causing agents to redundantly re-read files, with each re-read appending yet another copy to the trajectory. To mitigate this, we propose CORVUS, a novel trajectory architecture that decouples file-read actions from their observations by maintaining a synchronized registry of relevant files and injecting only their current contents at each reasoning cycle. This structural change produces significantly lighter-weight trajectories that remain synchronized with the actual codebase state by construction, eliminating redundant file copies and stale snapshots that bloat conventional trajectories. We evaluated CORVUS on SWE- POLYBENCH_VERIFIED and SWE-BENCH PRO across four LLMs, achieving 9-50% reduction in average input tokens per task, 15-32% shorter final prompts, and up to 37% fewer reasoning cycles while maintaining comparable pass rates.
Read the full abstract

All authors Mingwei Zheng, David OBrien, Siwei Cui, Pardis Pashakhanloo, Rajdeep Mukherjee, Myeongsoo Kim, Sachit Kuhar

Machine Learning

Measuring the Dependency Gap: Diagnosing Inter-Column Fidelity in Tabular Generative Models

Jie Zhang

Synthetic tabular data are valued for preserving inter-column dependency, yet a routine fidelity score says neither where it is lost nor why. Equipping a classifier two-sample test (C2ST) with a gradient-boosted discriminator, we decompose it by controlled permutation into marginal, dependency and numerical–categorical cross components, read against a fully-factorized reference that destroys all dependency while keeping every marginal, and a real-data oracle. The linear detection score in common use rates that reference as nearly real, a known weakness we replicate on four benchmarks. On a flow-matching (TabbyFlow) and a diffusion (TabDiff) generator it exposes a deficit of the same order in both, whose source we narrow by elimination: not a structural limit of the mean-field objective, not mainly sampling discretization, and not closed by a $16\times$ capacity increase where that test has power, while the same measurement responds sharply when capacity is cut eightfold, so the plateau is measured rather than a blind spot. Destroying inter-column dependence while keeping every marginal and every feature–label relation costs between $0$ and $0.22$ minority-class $F_1$ by dataset, and the generators' own shortfalls are not ordered by their deficits. An explicit cross-coupling module and a post-hoc copula leave the deficit in place. Neither generator's objective places a term on the joint at all; that this absence is what sustains the deficit is an explanation we reach by elimination rather than test. A controlled validation calibrates the diagnostic and identifies when it reports a false positive, which none of these generators triggers.
Read the full abstract

Machine Learning

Toward User-Conditioned Evaluation of Personal LLM Agents under Temporal Interventions

Pin Qian, Su Wang, Yihang Chen et al. (+5 more)

Personal agents maintain memories, learned skills, tool configurations, and policy state that evolve with each user. Existing agent benchmarks often evaluate these capabilities in isolation: tool benchmarks test invocation under fixed APIs, memory benchmarks test recall or forgetting, and safety benchmarks test static policy compliance. We argue that personal-agent evaluation requires a different protocol: replaying the same temporal intervention across different persistent user-conditioned states and measuring how failures propagate across agent components. We formalize this requirement as four conditions: explicit temporal intervention, persistent state across the intervention, induced cross-dimensional effects, and variation in user-conditioned state. A focused audit of public benchmark protocols selected by explicit inclusion criteria identifies several close cases. Under our explicitly narrow operationalization, we did not find a protocol in that audited set satisfying all four conditions. This claim is scoped as a focused gap analysis with bounded literature coverage. This position paper proposes a minimal benchmark design and candidate reporting metrics for user-conditioned adaptation. The result is a concrete design requirement for future personal-agent evaluation, with metrics used as reporting tools for that requirement.
Read the full abstract

All authors Pin Qian, Su Wang, Yihang Chen, Qiaolin Yu, Xiaoyuan Wang, Zhitong Guo, Zhicheng Wang, Junxian You

Machine Learning

MotifRole-Diff: Risk-Optimal Role-Aware Corruption for Masked Molecular Graph Diffusion

Tasfia Nuzhat Ornee, Elias Hossain, Ivan Garibay et al. (+1 more)

Masked discrete diffusion for molecular graph generation typically applies a uniform corruption schedule to all tokens in a lossless graph-to-sequence representation, implicitly treating structurally heterogeneous molecular components as equally difficult and equally important to reconstruct. However, different molecular graph token roles exhibit substantial variation in denoising difficulty and their influence on the decoded molecule, motivating role-specific corruption strategies. We introduce MotifRole-Diff, a role-aware corruption process that allocates masking rates according to empirically measured denoising difficulty and graph-level perturbation impact while preserving the model architecture, clean sequence space, and lossless molecular-graph decoder. We formulate schedule selection as the risk-optimal allocation of a fixed masking budget across token roles. Our theorem characterizes optimality for the modeled role-weighted residual risk, while downstream generation performance is evaluated empirically. Under matched architecture, training budget, and sampling compute, MotifRole-Diff improves validity on QM9 from 0.905 to 0.944 while reducing FCD from 1.701 to 1.609, and on MOSES improves validity from 0.920 to 0.938 while reducing FCD from 2.125 to 1.850. Role-wise diagnostics further show improved reconstruction across molecular graph token categories. Together, these matched-compute results indicate that structurally informed corruption is a more effective masking strategy than uniform schedules for serialized molecular graph diffusion.
Read the full abstract

All authors Tasfia Nuzhat Ornee, Elias Hossain, Ivan Garibay, Niloofar Yousef