Papers

The widespread adoption of high-level deep learning libraries, while accelerating model development, has increasingly abstracted away the internal mechanics of neural networks, creating a gap between practical usage and fundamental understanding. To address this, the paper presents a self-contained neural network framework implemented entirely from scratch without relying on automatic differentiation or pre-built deep learning modules. The implementation encompasses all essential components, including multi-layer architectures, diverse activation functions, regularization techniques, and state-of-the-art optimizers. Beyond serving as a pedagogical instrument that demystifies forward/backward propagation, gradient dynamics, and optimization landscapes, the framework demonstrates robust performance when applied to a multi-class classification task, successfully validating its correctness, numerical stability, and generalization across varied configurations. The extensible design and clean modularity further position it as a reliable baseline for educational purposes and future research exploration.

Early sepsis prediction from electronic health records is challenged by irregular sampling, high missingness, and class imbalance. We systematically compare four modeling paradigms – self-supervised Joint Embedding Predictive Architecture (JEPA) via masked latent prediction, self-supervised VICReg (variance-invariance-covariance regularization) with two-view augmentation, semi-supervised fine-tuning of a VICReg-pretrained encoder, and supervised Temporal Convolutional Network (TCN) – alongside raw-feature baselines. All models share a common preprocessing pipeline of hourly binning with forward-fill imputation applied to 7 biomarkers selected via sparsity analysis from the MIMIC-III dataset. Our best model (JEPA + XGBoost + mean pooling) achieves AUPRC 0.636 at the time of onset (H0), approaching the SupMix benchmark (0.667) while using 83% fewer biomarkers. The Tier 1 pipeline – VICReg pretraining followed by semi-supervised fine-tuning and XGBoost – achieves AUPRC 0.510 at H0, a 3.1$\times$ improvement over the raw-feature baseline (0.165) and a 7.6% improvement over the end-to-end supervised TCN (0.474). Crucially, the fine-tuned VICReg encoder exhibits the most temporally persistent representations, degrading only 16.8% from H0 to H10 compared to 47.5% for supervised TCN and 65.3% for JEPA, demonstrating that self-supervised pretraining with task-aware fine-tuning yields features that are both sharp near onset and robust across prediction horizons.

How deep does a graph neural network need to be on a sparse graph? We study its purest statistical form: node classification on the sparse contextual stochastic block model (CSBM) with average degree $\Delta=O(1)$, whose local weak limit is a broadcast-labelled Poisson Galton-Watson tree. Prior work derived a message-passing classifier $h_\ell$ that aggregates from each vertex at distance $k\le\ell$ the attenuated evidence $2\operatorname{artanh}(\gamma^k t(X_v))$, with $\gamma$ the edge signal and $t$ a bounded likelihood-ratio transform of the feature. We prove that the value of depth is governed by a single number, the Kesten-Stigum ratio $\kappa=\gamma^2\Delta$. Below the threshold ($\kappa<1$), the error sequence is Cauchy at a geometric rate, $|\mathcal{E}(\ell)-\mathcal{E}(\ell')|\le C\kappa^{(\ell+1)/3}$ for all $\ell'>\ell$, so all layers beyond depth $O(\log(1/\epsilon))$ change the error by less than $\epsilon$; conversely, under mild regularity each sufficiently deep layer still flips the decision with probability at least $c\kappa^{\ell/2}$, the empirically sharp exponent. Above the threshold ($\kappa>1$), depth is geometrically productive: $\mathcal{E}(\ell)$ is driven to a branching-process floor of order at most $1/(\kappa-1)$ at any geometric rate $\kappa^{-s\ell}$, $s<1$ (this bound has content only for $\kappa>17$). No local classifier of any depth beats the universal floor $e^{-\Delta}\Phi(-\zeta)$ set by isolated roots ($\zeta$ the feature signal-to-noise ratio), while the first layer provably helps by an explicit total-variation amount. Simulations with an exact belief-propagation baseline on the same trees show that the pairwise rule's error curve is mildly non-monotone in $\ell$, so an optimal finite depth exists (an exact instance is certified in the appendix), while BP saturates strictly faster, at an effective per-layer ratio below $\kappa$ that we identify.

A predictor with small overall bias can still exhibit systematic bias conditional on its predicted value, limiting its reliability for downstream uncertainty quantification and candidate screening. We propose Isotonic Conformal Inference, a unified framework for prediction and multiple testing that separates calibration from inference and exploits the level set structure of an isotonic predictor. This structure enables the computation of localized conformal quantiles and $p$-values while avoiding repeated isotonic refitting across candidate responses. For prediction, we introduce Split Isotonic Conformal Prediction (sICP) and Transductive Isotonic Conformal Prediction (tICP), establish prediction-conditional coverage guarantees for both methods, and characterize their finite-sample and asymptotic self-calibration properties. For selection, we introduce Isotonic Conformal Selection (ICS), which provides finite-sample stratum-conditional false discovery rate control. Experiments on simulated regression problems, a healthcare-utilization dataset, and a high-throughput ADMET drug discovery dataset demonstrate the practical performance of the framework and substantially lower computational costs relative to full transductive baselines.

Accurate routability estimation during physical design is important for reducing costly post-routing iterations. Prior learning-based methods treat this task as deterministic prediction, mapping placement-stage features to a single congestion or DRC outcome. We instead formulate routability estimation as a conditional generation problem, where both routing congestion and DRC violations are modeled as spatially structured routability fields. Our framework, Conditional Latent Diffusion for Routeability estimation (CLDRoute), uses physics-aware conditioning and task-specific latent modeling to handle the different characteristics of congestion and DRC maps. This allows our method to supports sample-based inference, producing both a mean prediction and a spatial uncertainty estimate for the same input design. On CircuitNet 2.0 (N28), our method achieves, for DRC violation generation, an SSIM of 0.9678, an MAE of 0.0028, and a TopK@1% of 0.3494; for congestion generation, it achieves an SSIM of 0.9031, an MAE of 0.0286, and an NZ-Pearson of 0.3692. Overall, our framework provides a more practical view of routability at placement by generating both the expected outcome and its uncertainty.

OpenLanguageModel (OLM) is an open-source PyTorch library for building and pretraining small language models while keeping their machinery visible. In OLM, model code reads like the architecture: components are ordinary modules, while Block, Residual, Repeat, and Parallel describe how they are wired. The resulting model can move unchanged from a teaching notebook to a complete pretraining run or a research ablation. OLM connects this readable model layer to tokenizers, local and streaming datasets, optimization, mixed precision, callbacks, checkpoints, and hardware-aware CPU, single-GPU, and single-node multi-GPU execution. We demonstrate the full path by tracing GPT-2 from diagram to code, launching a FineWeb-Edu training script, replacing one attention component, and letting AutoTrainer configure the available machine. The package includes 27 presets across nine familiar model families and documentation that progresses from LM fundamentals to architecture research. Validation shows close agreement with independent reference implementations, 90.6% four-GPU weak-scaling efficiency for a 348M-parameter workload, compact architecture edits, and positive early usability results. OLM is MIT-licensed and available through PyPI, GitHub, and its documentation site.

Accurate emergency triage decision is critical to avoid clinical deterioration, morbidity, and mortality. Machine learning-based triage system involves acquiring the main presenting complaint in text form and assessing vital signs in numerical data, enabling an automated and efficient analysis of patient information for timely and accurate prioritization of medical attention. However, modelling the intricacies of both data types requires a comprehensive understanding of the temporal structure and dependencies within the data. Thus, the aim of this study is to propose a multimodal deep learning architecture that can effectively handle both tabular and textual data. Furthermore, the proposed model exploits self-attention to to capture both local and global relationships between the features. A dataset consisting of 11,102 triage data collected from emergency department of Hospital Universiti Sains Malaysia is used for model development and validation. The proposed model demonstrated an increase of 1.95% in accuracy, 2.49% in F1-score, and 1.41% in ROC AUC compared to the baseline model. The experimental results demonstrated the potential of the proposed model in predicting triage decisions.

The increasing adoption of Large Language Models (LLMs) as AI components in modern software systems introduces distinct security risks to the software supply chain. While many considerations and safety mechanisms are in place for components of the traditional software supply chain, the recent rapid adoption of AI components and platforms has overlooked these hard learned lessons. Selecting and integrating AI models without clear guidance on how these choices affect system security may leave applications vulnerable to threats, such as malicious components, data leakage, and unintended behavior. The goal of this study is to understand practitioners' decision making process and security considerations in selecting and integrating AI components through an exploratory semi-structured interview study. Toward this goal, we conducted semistructured interviews with 22 software developers, architects, and AI practitioners across diverse organizations about how they integrate AI components into their software. Our analysis finds that practitioners' model selection is predominantly driven by functional criteria, including performance, accuracy, cost, and specific features, e.g., tool calling or multimodal support, while security is rarely considered as an evaluation criterion. We observe a consistent lack of security concern throughout the AI component integration process, with established software supply chain lessons overlooked or ignored. The industry is repeating the historically costly mistakes of early software dependency management, prioritizing rapid reuse and availability over security and provenance. We distill our findings into actionable recommendations for AI adopters, model providers, and researchers, advocating for a proactive, security-by-design approach that integrates security evaluation into component selection and sustains it throughout the software development lifecycle.

Principal component regression (PCR) regularizes high-dimensional prediction by choosing a spectral cutoff, but rank selection cannot correct systematic inflation of the retained empirical eigenvalues. We study clean Gaussian random designs in which the aggregate covariance tail creates a nearly scalar sample-space floor comparable to the predictive head scale. De-floored principal component regression (dPCR) retains the cutoff and subtracts an estimated floor from the retained denominators. We prove an ordinary-PCR prediction-risk lower bound uniform over all ranks and a high-probability dPCR upper bound. When the floor is sharp and inexpensive to remove in population prediction risk, the conditional risk of dPCR is asymptotically negligible relative to that of the best ordinary PCR rank. An exact risk decomposition explains the separation: denominator inflation is governed by first spectral mass, whereas the clean prediction cost of correction is governed by squared spectral mass. A same-sample trimmed-mean floor estimate attains the oracle dPCR upper-bound rate at a prespecified rank, and the separation persists under approximate predictive alignment when the tail prediction-energy fraction vanishes. Separate pointwise fixed-aspect formulas show that the risk-optimal positive scalar correction improves rank-$1$ PCR, whereas mean-floor subtraction is generally not optimal for a broad Marchenko–Pastur bulk.

Differential privacy (DP) is increasingly deployed to limit membership inference risk in machine-learning systems. Prior work has shown that DP-SGD can widen accuracy disparities across demographic groups, but this framing treats fairness as a purely outcome-side concern. We argue that privacy cost, the information leakage borne by each group, is itself a form of harm, and adopt a compensatory-fairness framework in which a group that involuntarily bears greater privacy exposure is owed proportionally greater benefit from the system. From this principle we derive the Privacy-Cost Equity Ratio (PCER), a group fairness metric defined as a group's positive prediction rate normalized by its per-group overfitting gap. By a standard membership inference bound, this overfitting gap upper-bounds each group's vulnerability to inference attacks, making PCER a conservative measure of benefit relative to exposure. PCER needs only per-group train and test accuracy (no shadow models), making it a practical post-hoc audit tool. We evaluate PCER alongside standard fairness metrics across six benchmark–attribute combinations spanning tabular and NLP domains, under DP-SGD at a range of privacy budgets, and validate the overfitting-gap proxy against a direct threshold membership-inference attack. The results reveal patterns that outcome-based metrics miss. On COMPAS, PCER uncovers a persistent double disadvantage: the protected group bears both greater privacy exposure and worse predictive outcomes, something demographic parity gap masks entirely. Sensitivity analysis shows very strong privacy guarantees collapse both groups' overfitting to a numerical floor, rendering exposure-based audits uninformative in that regime. Together, these findings show that fairness audits of privacy-preserving systems must account for who bears the cost of protection, not only who benefits from its outcomes.

Backpropagation makes training deep networks memory intensive because it must store intermediate activations. Forward-mode methods avoid this cost, but their gradient estimates become increasingly noisy as the number of trained parameters grows. We introduce Split Forward Gradient (Split-FG), which splits a network at an intermediate representation: it computes the output head gradient exactly and estimates only the trunk gradient with a Jacobian–vector product. This reduces estimator variance and requires no backward pass through the trunk, while retaining an Adam-style convergence guarantee. Our experiments reveal an important practical failure mode. On WikiText-103, naive forward-gradient training of the trunk performs worse than leaving a randomly initialized trunk frozen, likely because Adam updates every noisy, under-determined trunk coordinate too aggressively. Simply using a much smaller learning rate for the trunk reverses this result: a $16$M-parameter GPT-2-style model reaches validation perplexity $387$, compared with $668$ for the frozen-trunk control and $2{,}885$ for a matched pure forward-gradient baseline (backpropagation reaches $150$). Split-FG also produces the strongest backprop-free results on our tabular benchmarks and reaches $60.5\%$ on CIFAR-10 and $35.2\%$ on CIFAR-100 with a heavy-head design. It reduces peak memory by up to $35\%$ relative to matched backpropagation, although the performance gap widens as the forward-mode trunk grows.

The RLxF programme argues that learning signals should come from world feedback rather than from internal model proxies. We instantiate this position in safe model-based control and distil it into three concrete design principles. Empirically, across four world-model architectures spanning a 2x MSE range, MPC planning is statistically equivalent (TOST, n=200), and dynamics-based uncertainty penalties increase collision rates from 26% to 34%: the standard MBRL safety proxy is anti-correlated with safety in this regime. Replacing the model-internal proxy with three world-feedback signals (a sensor-derived margin via minimum lidar, a temporal signal via time-to-collision, and an outcome-supervised feedback model g_psi trained on prior collision labels, structurally analogous to outcome-trained reward models in RLHF) reduces collisions to 1-14% without retraining the world model or the planner. The mechanism is structural: model uncertainty has support over state-prediction space, whereas task risk has support over constraint boundaries, with empirical correlation r < 0.15. From this we extract three RLxF principles (ground risk in world outcomes, validate proxies before deployment, and substitute outcome-trained feedback models when direct world signals are unavailable) and argue they apply equally to model-based control and to verifier-based or RLHF approaches in LLM alignment.

Kirigami turns stiff sheets into compliant, shape-morphing structures, but its reliance on periodic cut patterns comes at a cost: correlated panel rotations couple extension to shear, so stretching one axis drives a parasitic shear that cannot be suppressed, and also confine anisotropic stiffness to a narrow, discrete set of responses that cannot be tuned independently. Biological tissues overcome an analogous constraint through controlled disorder, such as graded fiber orientations in skin and hierarchical anisotropy in myocardium, achieving direction-dependent mechanics unavailable to regular architectures. Here, we show that engineered disorder is a design degree of freedom for kirigami, with stochastic kirigami accessing a continuous and far broader region of mechanical response than periodic patterns. This includes programmable anisotropy with near-complete elimination of extension-shear coupling. Because disordered patterns lack a simple parameterization, we navigate this design space with a geometry-aware graph neural network (GNN) that maps cut topology to the full nonlinear, bidirectional stress-strain response, coupled to a genetic algorithm that inverse-designs patterns reproducing target responses along two perpendicular axes. The GNN trains an order of magnitude faster and more accurately than image-based models. Fabricated elastomer samples reproduce the predicted nonlinear, anisotropic responses, closing the loop from design to physical component. By turning disorder into a variable to control directional stiffness, this work develops architected materials that stretch without parasitic shear, from soft actuators to tissue-interfacing devices matched to the anisotropy of living tissue.

Organic mixed ionic electronic conductors (OMIECs) are a promising class of polymer materials for applications spanning neuromorphic computation to energy efficient electronics and bioelectronics. Despite being highly tunable, the relationship between structural features and key performance properties such as charge carrier mobility is poorly understood. Scanning nanodiffraction in the transmission electron microscope (TEM) is a powerful probe for elucidating this structure-property relationship, but produces large, noisy datasets that are difficult to interpret because polymer reflections exhibit several distinct morphologies. To address the complexity, we trained a machine learning (ML) model to detect these polymer diffraction peaks and their intensities from synthetic data. Compared to correlative peak detection algorithms, the conventional method for analyzing nanobeam 4D scanning transmission electron microscopy (4DSTEM) data, we show that the ML model is significantly faster and outperforms correlative algorithms in almost all cases, opening up the possibility of near-live visualization of 4DSTEM experiments.

Function-preserving network growth techniques such as Net2Net and progressive stacking expand a model's capacity without destroying its learned function, but existing formulations either tolerate numerical perturbations or require a full rebuild of the training program. We formalize Exact Network Surgery: the in-place insertion of a residual block into a live computational graph such that (i) the network function is preserved – bit-exactly under explicit floating-point hypotheses – and (ii) inserted parameters remain trainable immediately after insertion. We prove an identity-morphism theorem for gated residual blocks, a structural-locality theorem showing that a reactive invalidation engine recomputes exactly the downstream cone of the insertion point, leaving every other node's value and optimizer state untouched, and an escape-from-initialization proposition showing that the Gradient Shadowing gate alpha, initialized at zero over a randomly initialized branch, receives a generically non-zero gradient at insertion time. We identify a degenerate configuration – zero-initialized output projections combined with a zero gate – that is an exact saddle point gradient descent cannot escape. Every claim is validated on the reference implementation in NeuroDSL, a reactive graph engine in Julia: grafting is bit-exact on every logit tested (0 mismatches out of 1600); the gate escapes zero at the first optimizer step and unlocks branch gradients at the second, exactly as predicted; the degenerate configuration exhibits gradients identically zero for the entire 600-step run; surgery cost tracks downstream cone size with r = 0.9992 while graft-plus-invalidation bookkeeping is constant (about 0.75 ms) across insertion depths; and training resumes bit-identically across a real process restart. A flagged preliminary appendix reports first single-seed observations on post-insertion gate dynamics.

We propose a language representation for multimodal data in which any observation, whether image, video, or text, is expressed as a bag of atomic propositions, simple statements about the entities, actions, and relations in a scene. A global semantic codebook unifies these into a shared vocabulary of canonical atomic propositions, placing every modality and observation into one interpretable space that spans fine grained facts to high level concepts and composes into richer ones. This brings interpretability with reasoning, cross-modal understanding and retrieval, and compositionality that enables complex multimodal understanding, rich data curation and complex structured retrieval. We demonstrate the framework on autonomous driving and open-world data.

Emergent capabilities are widely treated as unpredictable: loss improves smoothly while abilities appear abruptly. Prior work offers early-warning indicators but never scores them as forecasts: no lead time at controlled false-alarm rate, no calibration, no negatives, no blind tests. We supply that discipline and show that, in grokking model systems and small language models, emergence timing is forecastable per run, in advance, with calibrated uncertainty. Across 30 transformers at identical configuration, the formation time of the previous-token head forecasts each seed's induction-head emergence at Spearman rho=0.977 with median lead 975 steps (~15% of training); a best-case loss rule ties the ranking with 50-step lead (a nowcast). Conformal intervals covered 15/15 held-out seeds, and the frozen rule passed blind pre-registered gates on TWO never-seen configurations (10/10 and 9/10 coverage). A trap-language rung then attacked our own rule as pre-registered: where previous-token context pays for the task itself, the bare precursor false-alarms on 10/10 capability-blocked runs, while the mechanism-composed conjunction is certified in both language classes (0 false alarms) and times emergence at rho=1.000. Finally, a gap-origin study broke the fixed offset (both lr and batch move the gap ~2.3x; no external clock owns it) and revealed the law beneath: across 80 valid-anchor runs the anchor fires at 0.843 of time-to-emergence – t_event ~= 1.19 x t_anchor – and this multiplicative rule passed its own blind gate (5/5) at a third unseen configuration. False alarms are certified against 33 manufactured negatives. The precursor leads across 3 public model families (Pythia, OLMo, OLMo-2; 7 suites), with OLMo-2 at 1B tokens showing precursor formed, capability absent. Four pre-registered kill criteria fired and are reported. Every freeze precedes its data in a public commit chain.

Multi-agent LLM systems are expected to be more reliable because agents can catch each other's mistakes. But peer pressure cuts both ways: the same correction that fixes a wrong answer can overturn a right one. The tempting safeguard is a brake that keeps the beneficial revisions and blocks the harmful ones. We show this brake is hard to build, for a simple reason: a revision is harmful exactly when the original answer was right, so deciding whether to block it is the same as knowing whether the model was already correct. This turns the open-ended hunt for a brake into one measurable quantity, the model's self-knowledge: any brake built from a deploy-time signal is a correctness probe in disguise, and self-knowledge is far from perfect (AUROC $\approx 0.64$–$0.89$ across six model families). We call this ceiling the wall. Even white-box steering of the model's own correctness direction does not breach it: it changes how often the model revises, but harmful and beneficial revisions move together. At population scale the wall becomes the cliff: when most agents start wrong, debate amplifies the shared mistake into a confident, wrong consensus. In our multiple-choice societies, more agents, more model diversity, and a stronger member do not fix it. What helps is adding information before the revision, not filtering after it. Local agreement is not global correctness.

Knowledge-editing benchmarks certify local correctness, whether an edited model produces the new fact on near-edit prompts but not how much of the original fact remains decodable inside the model. We study residual knowledge directly with a linear trace probe: after editing a fact, we ask whether the original object is still recoverable from the model's hidden states. On GPT-2-XL, across three mechanistically distinct editors applied to 50 CounterFact edits, the original object remains linearly decodable well above chance after a successful edit (probe accuracy 0.96 for ROME, 0.86 for constrained fine-tuning, and 0.79 for the memory-based editor GRACE, against a chance level of 0.50; all edits reach 100% generation-based success). The GRACE result is the most informative: GRACE changes zero base-model weights, overriding the fact through an external memory, yet the original object is still decodable from the underlying network, so the residual trace cannot be attributed to an incomplete weight update. We read this as evidence that editing, even when behaviorally successful, suppresses rather than erases the original association in representational space. We also report a relearning-savings instrument that did not behave reliably in our setting and discuss why; we treat it as a negative methodological result rather than evidence. Code and data are released.

Inertial sensors at multiple body locations can improve activity recognition, but requiring every sensor at inference increases the deployment burden. We study whether four synchronized IMUs available during training can improve a student that uses only the right-arm IMU during fitting and inference. A frozen four-IMU teacher provides logit and feature targets. Fixed-weight knowledge distillation applies each target with the same strength to every fitting sample, although the student may not benefit equally from them. We introduce dynamic influence weighting (DIW), which tests a one-step candidate update on separate fold-internal training participants. DIW then assigns separate sample-wise gates to the logit and feature losses. On WEAR, we evaluate 19 labels and 68,298 complete windows from 22 participants using subject-disjoint five-fold cross-validation. Pooled out-of-fold macro-F1 is 0.561820 for Supervised and 0.571623 for Fixed-weight KD. DIW reaches 0.638451, gains of 7.66 and 6.68 percentage points, respectively. It exceeds Supervised for 18 of 19 labels and 21 of 22 held-out participants. All three routes retain the same 80,915-parameter right-arm student at inference. Under this protocol, DIW converts training-only multi-position information into a stronger single-IMU model without changing deployed sensing or the student forward graph.

Wafer fabrication exhibits unique characteristics, including reentrant process flows, variable bottlenecks, and highly variable process conditions. In order to identify the most severe bottleneck at each moment in time for semiconductor wafer fabrication, this research presents the Dynamic Multi-Criteria Bottleneck Severity Index (DMBSI), a new, data-driven methodology for analysing multiple diagnostic signals of cycle time to changes in process parameters, and the impact of reworks on cycle time in order to generate an interpretable, unified measure of bottleneck severity. The experimental validation of DMBSI was conducted using Manufacturing Execution System (MES) logs collected from 22 wafer production lots at a commercial 200 mm wafer fabrication line operated by Seagate Technology. Using 5-fold cross-validation, the GA-optimised DMBSI achieves a Pearson correlation of r = 0.80 with observed cycle-time contributions, representing an 8.1% improvement over the expert heuristic baseline (r = 0.74) and substantially outperforming the Theory of Constraints (TOC; r = 0.60) and Value Stream Mapping (VSM; r = -0.30). Furthermore, the unique time-windowed component of DMBSI enabled the identification of temporal bottleneck migration patterns, which shifted from the dominant constraints associated with dielectric deposition steps in the early production windows to those associated with over- and under-inspection in the later production windows. The integrated what-if counterfactual analysis demonstrated that a 50% reduction in waiting time at the top-ranked bottleneck step would reduce the mean cycle time by 7.2%, with the top five bottleneck steps offering a combined potential reduction of approximately 19%.

Retrieval-Augmented Generation (RAG) systems retrieve and integrate external knowledge to ground large language model (LLM) outputs. However, current RAG architectures treat all retrieved facts as equally valid regardless of temporal provenance, leading to temporal hallucination, where plausible but obsolete facts corrupt the output. A clinical lab reading from yesterday is actionable; the same reading from six months ago is noise. We present Chronofy, a three-layer neuro-symbolic framework implementing the Temporal-Logical Decay Architecture (TLDA) that embeds temporal validity directly into the representation, retrieval, and reasoning layers of RAG systems. Layer 1 reserves a dedicated temporal subspace within Matryoshka embeddings to make fact age structurally irremovable from the representation. Layer 2 integrates learnable exponential decay functions into graph-based retrieval, where the decay coefficient $\beta_j$ is grounded in Bayesian decision theory as an approximation of twice the latent process mean-reversion rate. Layer 3 applies Signal Temporal Logic (STL) robustness functions to evaluate the temporal validity of retrieved knowledge, not LLM output confidence, and enforces the possibilistic weakest-link principle to bound output confidence by the most decayed evidence in the reasoning chain. We evaluate Chronofy on temporal knowledge graph forecasting benchmarks, the TimE temporal QA benchmark, and a domain-specific sensitivity analysis, demonstrating that explicit temporal decay modeling improves retrieval precision, reduces temporal hallucination, and enables principled data re-acquisition triggers when temporal context is insufficient.

Monitoring poverty and food security indicators is imperative for addressing socioeconomic challenges in developing nations. A limitation is mismatches in scale between data sources: census data provide geographic coverage, while socioeconomic indicators are derived from infrequently conducted surveys at coarse resolutions, posing a methodological challenge. This study introduces a deep learning framework, JuGAAD, using Indian census and survey data from 2001 and 2011 as a case study. We employ a three-step process: census and geospatial data are averaged into intermediate village-cluster-scale tessellations to reduce noise and regularize administrative boundary changes; an autoencoder compresses high-dimensional National Sample Survey Office (NSSO) data into a low-dimensional latent representation; and a regression model maps upscaled census and geospatial data to this representation. This function is applied to fine-grained census data to generate high-resolution predictions, validated against ground-truth district-level NSSO indicators. Results confirm the methodology predicts socioeconomic indicators at fine scales with strong accuracy.

AI Assistants are increasingly deployed in high-stakes settings, such as healthcare or government services. Yet their real-world behavior remains poorly understood due to strong context dependence. Current evaluation protocols follow a defense-in-depth paradigm with compounding layers of safeguards, ranging from traditional benchmarks to live or adversarial testing. Such benchmarks remain largely static and single-turn, limiting their ability to capture real-world variability in conversational settings. We propose StabilityBench, a principled, general and model-agnostic benchmark operator that transforms single-turn benchmark queries into multi-turn interaction histories. StabilityBench augments existing benchmarks by injecting realistic user simulations, through demographic proxies or sycophantic baits, while preserving original task intent. We apply StabilityBench to four benchmarks spanning mathematical reasoning, health question-answering and safety, and evaluate nine large language models under these conditions. Our results show that model performance is consistently unstable under these injections, with considerable performance degradations on three out of four benchmarks studied. These highlight important limitations of static evaluations and motivate more realistic evaluation settings. To this end, we propose StabilityBench-Mini: a size-preserving variant of StabilityBench that samples across diversification axes, enabling more realistic evaluation without increasing costs.

Scientific discovery is increasingly shifting from isolated disciplines to multi-domain reasoning, and AI for science faces a similar transition. Existing systems are either specialised for individual domains or unify scientific data mainly through text tokenisation and prompt-based interfaces, limiting their ability to handle diverse scientific inputs, produce modality-native outputs, and support joint understanding, reasoning, and generation across scientific domains. We introduce MKB, a unified scientific multimodal model for both understanding and generation, built around a shared Transformer backbone and modality-tailored encoders, adapters, and decoders. MKB covers six scientific branches, including DNA, RNA, proteins, small molecules, earth science, and medical images, and supports native outputs such as biological sequences, molecular strings, meteorological fields, and segmentation masks. Training follows a two-stage modality-then-language curriculum: Stage 1 aligns modality-specific components with the frozen backbone, and Stage 2 consolidates them with the language backbone using mixed scientific and general corpora. Experiments show that MKB achieves competitive scientific understanding across biological and molecular benchmarks, produces high-fidelity native outputs for weather forecasting, biological generation, and medical-image segmentation, and largely retains the general capabilities of its Qwen3-VL backbone. These results demonstrate the feasibility of the proposed paradigm, suggesting that shared-backbone models with modality-tailored components can provide a promising foundation for future cross-domain scientific multimodal exploration. The model and code are publicly available at https://github.com/Shanghai-Academy-of-AI-For-Science/MKB and https://huggingface.co/sais-org/MKB.

Experiment trackers show how training is progressing, but changing a live run still usually requires trainer-specific code. We present Interactive Training 2, an open-source control plane for steering training through a shared protocol. Training applications declare which settings and actions they expose, humans and automated controllers submit requests through the same interface, and the training loop validates and applies them at safe control points. A customized Aim workspace combines live metrics and controls with a chronological record of requests and outcomes. We demonstrate the system across five NLP and reinforcement-learning workflows. The released code and traces provide a reusable foundation for auditable human- and agent-guided training.

Comparing phylogenetic tree topologies is essential for understanding epidemic dynamics, yet biologically meaningful distances such as the Subtree Prune and Regraft (SPR) distance are NP-hard to compute and intractable on large datasets. We investigate whether a Graph Neural Network (GNN) can approximate SPR distances in near-constant time per comparison after training. Our contributions are fourfold. First, we build and publicly release a dataset of 864 phylogenetic trees inferred with UPGMA and Neighbor-Joining over four bacterial species, spanning up to 9{,}500 isolates, together with 388 labelled tree pairs. Second, we establish a reproducible pre-processing pipeline including midpoint re-rooting, which reduces tree depth and supplies the rooting required for exact distance computation and for the model's root-based features. Third, we validate the supervision target: on small trees, where exact SPR is tractable, the unrooted phangorn::SPR.dist heuristic correlates almost perfectly with the exact rooted distance computed by rspr (Pearson $0.98$–$0.99$), making it an excellent monotonic surrogate. Lastly, we train a Siamese Graph Isomorphism Network (GIN) regressor. In-distribution, i.e., held-out trees from the same species and size range as training, it explains roughly 87–90% of the variance ($R^2 \approx 0.87$ on a held-out split; $0.90 \pm 0.19$ under stratified cross-validation), with about four times lower error than a mean-predictor baseline, and shows partial transfer to unseen species ($R^2 \approx 0.37$). Its main limitation is extrapolation to trees larger than those seen in training, where accuracy collapses. The released dataset and the validated heuristic versus exact relationship provide a reproducible basis for scaling learned SPR approximation.

Unsteady aerodynamic phenomena, such as gusts, turbulence, and fluid-structure interactions affect an aircraft during flight. For design, optimisation and certification, it is indispensable to quantify such unsteady aerodynamic effects. Industry-standard computational fluid dynamics methods, such as solving the unsteady Reynolds-averaged Navier-Stokes equations or the linearized frequency domain method, are either computationally expensive or restricted by assumptions like linearity. Once trained, machine learning methods are capable of computing non-linear relationships very fast, making them suitable as surrogate models. By autoregressively applying graph neural networks (GNNs), operating on a discretised spatial domain, spatio-temporal predictions can be made. However, autoregressive GNNs suffer from error accumulation leading to unstable rollouts over time. Here we show that combining GNNs with augmented Neural Ordinary Differential Equations yields temporally stable predictions of the surface forces on a pitching airfoil. We found that our approach, called GNODE, based on Graph Neural Ordinary Differential Equations, provides temporally more stable, spatially smoother, and overall more accurate results than an autoregressive GNN baseline. Tests are conducted on a dataset consisting of a simulations of a pitching airfoil, including transonic shocks, transient behaviour and dynamic non-linearities. Augmenting GNODEs with additional latent dimensions improves the expressivity and accuracy by capturing underlying history effects. The developed method demonstrates an approach that is suitable to model non-linear spatio-temporal systems with exogenous inputs.

Calibration of grey-box simulation models is a constrained optimization problem in which model evaluations are expensive, the parameter space can be high-dimensional, and the search must respect plausibility constraints. Although the simulation code is fully available to the analyst, the joint effect of multiple parameters remains difficult to predict analytically. Classical optimizers such as Nelder–Mead (NM) are simple to deploy but sample-inefficient, particularly under constraints. Modern Bayesian Optimization methods achieve competitive solutions with far fewer evaluations but require non-trivial modeling machinery for constraint handling. We introduce an agentic calibration method in which a large language model acts as the optimizer, with constraints incorporated as a plain-language section of the system prompt. We evaluate the agentic method, NM, and Bayesian Optimization (BO) on an anal cancer simulation model under both unconstrained and clinically constrained calibration. Under unconstrained calibration, the agentic method achieves substantially lower best error than BO and NM, while requiring fewer model evaluations. Under constrained calibration, the agentic method reaches comparable error levels and both outperform NM. These results are obtained at the cost of increased inference time per iteration. Agentic calibration achieves competitive performance with substantially fewer model evaluations, and constraint handling is essentially free at the modeller-facing interface through simple textual specifications rather than additional modelling machinery. The main trade-off lies in increased per-iteration inference cost, making the approach particularly suitable when simulation time dominates. Beyond performance, the per-iteration rationale makes the search auditable and explainable, so its decisions can be scrutinised and justified to third parties.

Sharpness-Aware Minimization (SAM) improves generalization by minimizing the worst-case loss in a local parameter neighborhood. Standard SAM implicitly allocates its global perturbation budget across parameter blocks according to instantaneous minibatch gradient norms. Such an allocation can be noisy and may not reflect the sensitivity that blocks accumulate throughout training. We propose Gradient-Energy Adaptive Radius SAM (GEAR-SAM), which maintains an exponential moving average (EMA) of squared block gradients as a lightweight, curvature-related sensitivity signal and allocates the fixed SAM budget through a closed-form constrained optimization. GEAR-SAM preserves the global SAM radius, requires no Hessian-vector products or explicit Fisher estimation, and adds only scalar state beyond SAM. Experiments on image classification, transfer learning, noisy-label learning, and partition studies demonstrate improved generalization and robustness across architectures and tasks. More broadly, GEAR-SAM provides a dynamic view of sharpness-aware optimization: a fixed perturbation budget should be redistributed as the sensitivity of functional network blocks evolves during training.