Vision-language models (VLMs) are increasingly used to make decisions from visual inputs. We introduce FAIRLENS, a benchmark and evaluation framework for measuring both the fairness and the validity of VLM responses in three high-stakes domains: hiring, legal, and healthcare. FAIRLENS pairs real face images spanning gender, race, and age groups with closed- and open-ended questions, giving more than 100K image-question pairs per model, and evaluates responses from four complementary views: demographic parity over adverse outcome rates, soundness, demographic association over unsupported roles and statuses, and bias in free-text generation. Soundness is the central validity criterion: a response is sound when it follows the evidence stated in the question and abstains when the image cannot support an answer. Evaluating eight VLMs, we find that the primary failure is unwarranted inference rather than unequal treatment. Models routinely infer qualifications, threat, illness, or professional role from a face instead of abstaining, and the weakest model does so on 99% of the questions its input cannot answer. These failures are most severe in legal and healthcare, where recognizing insufficient evidence matters most, and disparity metrics alone would miss them: parity gaps are small in absolute terms, yet when baseline adverse rates are low the same gap means one demographic group receives adverse labels several times as often as another, and a small gap can equally reflect a model that treats every group unsafely. Bias in free-text responses is only loosely coupled to multiple-choice accuracy, so correct structured answers do not imply safe generation. FAIRLENS shows that fair high-stakes VLM behavior requires similar treatment across groups and refusal to infer high-stakes attributes from appearance, and its question suite transfers to any face corpus with demographic annotations.
Vahid Reza Khazaie, Ahmed Y. Radwan, Shaina Raza· 0 citations
Vision models deployed on microcontrollers (MCUs) are quantized to integer-only arithmetic and run in inference-only runtimes that do not carry the machinery backpropagation needs: the standard tool for adapting a model to the distribution shift (sensor noise, blur, lighting) it meets in the field. Existing forward-only test-time adaptation (TTA) methods either run only on server- or edge-GPU-class models (not true microcontroller integer execution), or require the batch-normalization (BN) layers that integer deployment fuses away. We present a forward-only TTA method that operates on deployed, BN-folded, integer-only convolutional networks. The key observation is that fusing BN into the preceding convolution, a mandatory step for integer inference, destroys the statistics that normalization-based adaptation relies on. We restore adaptation by re-normalizing each folded convolution's per-channel output to its clean training statistics, using only forward-pass estimates. The method (i) recovers most of gradient-based TENT's accuracy gain (+20.9 vs. +24.9 points) and matches forward-only BN adaptation, while being the only method that runs on a folded integer-only model; (ii) needs to adapt only 3 of 21 layers (selected without seeing the test corruptions) to recover 93% of the benefit; (iii) survives single-sample streaming with a batch-size-scaled momentum; and (iv) generalizes across three datasets (up to 200 classes) and two architectures. We validate bit-exact int8 convolution execution and deploy on an ESP32-S3, where, measured with a Nordic PPK2 power profiler, the forward-only adaptation (a lightweight fp32 recalibration around the int8 convolutions) costs only 8.3 mJ (6.8% of inference energy) and 21.9 ms on the deployed SIMD-optimized model: forward-only adaptation is cheap on a real microcontroller.
Muhammad Rehan, Haider Ali, Muhammad Ali Munir et al.· 0 citations
Accurate large-scale wildfire spread modelling requires models that capture both the environmental conditions associated with fire occurrence and the local dynamics of fire propagation. We propose a three-stage framework that combines a Random Forest (RF) model with a cellular automaton (CA). First, an RF model trained on the 2021 Canadian fire season estimates daily pixel-level fire-occurrence probabilities. Second, quantile gradient boosting models provide optional spread-rate priors for sensitivity analysis. Third, an RF-informed CA combines the RF probability layer with neighbourhood-driven spread on a 5 km grid. The RF model achieved AUC values of 0.725--0.795 on the 2022--2024 datasets, while the RF-informed CA achieved substantially higher spatial overlap than the evaluated CA-only baselines in the 2023 simulation. A higher-resolution simulation provides an additional qualitative assessment of local spatial errors. These results suggest that combining RF-derived probabilities with local CA spread can improve large-scale wildfire simulations under the tested conditions.
Artificial Intelligence has shown to help improve agricultural practices, yet adoption remains limited: 69% of U.S. farmers have privacy concerns with sharing their data, and these concerns must be addressed before adoption is widespread. While Federated Learning has been demonstrated to protect privacy at scale for other sectors, deploying a system for agriculture comes with its own set of challenges; the problem necessitates a system that can protect farmer data and identities while preserving model utility, runs on commodity hardware, and is resilient to fragile rural infrastructure. To address these concerns, we introduce the Private Computation Space (PCS), a deployed, open-source Machine Learning system to provision and process farmer data securely. We design a system tailored to an agricultural setting, with multi-cluster orchestration for reliability in rural areas with asynchronous Federated Learning (FL), Differential Privacy (DP), and Trusted Execution Environments (TEEs), to allow farms to participate in the framework while keeping their data private. We evaluate the system on two deployed workloads: monitoring nitrogen with living plant sensors in NY for six months and predicting evapotranspiration from weather stations in CA for ten months. Our evaluation finds a Dice Similarity Coefficient (DSC) of 0.71 and $R^2$ accuracy of 0.84 for the respective workloads, improving the worst single-site model accuracy by 22.4% and 9.1%, respectively, while preserving privacy.
Shuang-Yu Lei, Muhammad Salman Abid, Jacob Belding et al.· 0 citations
Reach audiences
Advertise in front of researchers, engineers, and readers.
Agentic AI systems are increasingly deployed to process sensitive data at inference time, such as healthcare records or financial documents assembled into a hidden \emph{context} before the system answers. Prior work has studied privacy risks primarily through \emph{jailbreaking} attacks that induce models to directly disclose sensitive content, but has largely overlooked the agentic setting where the context is assembled by the agent's own tool calls. We show that the agents we evaluate remain vulnerable to hidden-context leakage despite the controls we test against them, namely an instruction not to disclose the context, logit suppression, and context dilution. For instance, a web-browsing agent answering benign user queries still carries exploitable signals about records silently loaded into its context. We introduce and formalize \emph{context-inference attacks} through a security game and evaluate three settings under decreasing attacker knowledge and increasingly indirect delivery of the context: a known context, an unknown context, and a context the agent retrieves through its own tool calls. We distinguish a grey-box setting, in which the target model is used to score observations, from black-box settings in which the attacker scores with a surrogate it controls. We further characterize how leakage varies with query budget, context size, and target-model size. A single attack carries through all three settings without modification, reaching $100\%$ ASR on small candidate sets and $63\%$ at $1024$ candidates against a known context, $78.9$ AUROC when the template and surrounding records are unknown, $92.5$ AUROC when a 14B surrogate scores a 32B target, and $81.8$ AUROC when the records arrive as an agent's retrieval returns, against chance rates of $1/|\mathcal{Z}|$ and $50$ respectively.
Prince Jha, Samuele Poppi, Nils Lukas· 0 citations
In the era of digital social interaction, searching friends' posts from massive social streams has become a fundamental user need. However, while modern agentic search frameworks have achieved remarkable success in conventional retrieval tasks, they break down when confronted with heterogeneous user queries and multi-dimensional social feeds, resulting in severe performance degradation in complex social search. To bridge this gap, we introduce SocialBuddy, the first agentic search framework tailored for social scenarios. Specifically, we construct SocialEnv, the first large-scale simulated environment for social search. Powered by an automated data and trajectory synthesis pipeline, SocialEnv includes 200K user profiles, 10 million social posts, and 50K reasoning trajectories, establishing a solid foundation for the development of social search agents. To tackle the credit assignment dilemma caused by sparse rewards in social search, we design SocialPO, a hybrid-granularity optimization framework. It macroscopically reinforces successful reasoning paths via multi-dimensional rewards, while microscopically rectifying deviated trajectories through fine-grained prefix truncation and token-level supervision. This hybrid-granularity design delivers multi-scale guidance in complex long-sequence scenarios. Finally, we construct SocialSearch Benchmark to provide a quantitative evaluation scheme for assessing the social search capabilities of SocialBuddy. Extensive experiments demonstrate that SocialBuddy-35B surpasses significantly larger frontier LLMs. Code and dataset will be released upon article acceptance.
Mingxuan Li, Yirong Mao, FaZhan Zhang et al.· 0 citations
A composite structural index summarises a network in one number; for a triangle-based index it is spectrally redundant: Tr(A^3) is the third moment of the adjacency spectrum. The non-redundant content sits one level down, in diag(A^3), which depends on eigenvectors and is not spectrally determined. A corollary in the theory paper for this index family stated that, and predicted: the global scalar should tie sharpened spectral baselines rather than beat them, while the node-wise attribution should do better where the number of structural epicentres is unknown. This paper tests it.
We construct Omega-N by localizing each of the four factors. The direct localization is badly conditioned; two corrections from published practice fix it, a configuration-null excess for every local factor and a personalized-PageRank neighbourhood at several scales, giving ten interpretable features per node from the graph alone, with no attributes, training or embeddings.
Against a recursive feature engine at five levels of recursion, Omega-N wins on one and ties on four of six in-domain node-classification evaluations, with ten features against up to 252. Two statistics computed from the graph and labels, not from performance, partition the eight benchmarks without error, and the two they exclude are the two it loses.
The strongest application is drug-target prioritisation on protein interaction networks: +0.073 to +0.144 AUPRC over a centrality battery across four constructions, replicated on an independent AP-MS network and label source, surviving three bias controls (degree-matched, ten repetitions: +0.1047 and +0.1030, both 10/10, p=0.00195). The clearest negative sits in the same application: adding Omega-N to centralities plus Node2Vec changes nothing (+0.0014, p=0.31). The claim is narrow: ten interpretable features
E-commerce search ranking must balance multiple objectives--relevance, user engagement, and platform revenue--when allocating impression slots to competing listings. Estimating the expected revenue component is well understood for fixed-price items, but becomes challenging when marketplace inventory includes mixed listing formats such as pure auctions and hybrid "Auction with Buy It Now" (ABIN) items, where prices evolve dynamically and the final transaction value is unknown at ranking time. Yet auction and ABIN listings account for a meaningful share of inventory and transaction volume on platforms such as eBay, and are a popular format for individual sellers and for unique items with unclear value. We extend the standard Expected Cost-per-Mille (eCPM) framework to auction and ABIN listings by deriving a marginal eCPM (meCPM) that captures the incremental value of showing one more impression of an item whose price is still evolving. The resulting formulation extends the familiar fixed-price eCPM--which is already inherently marginal--to auction dynamics, allowing unified ranking of fixed-price, auction, and ABIN listings under a single objective. We then describe a practical production implementation that approximates this objective, addressing cold-start challenges by bootstrapping from existing engagement models. Online A/B tests at a large e-commerce platform showed positive revenue gains and statistically significant improvements to user metrics, and the system was deployed to production.
Autonomous and intelligent transportation systems operate in complex urban environments where safety depends on interactions among vehicle behavior, environmental conditions, and vulnerable road users (VRUs) such as pedestrians and cyclists. Most advanced driver assistance systems (ADAS) employ reactive mechanisms that activate only after hazards have emerged, a critical limitation underscored by rising VRU fatalities in the United States.
This study introduces PRISM (Proactive Risk Intelligence and Safety Management), an agentic multi-model safety architecture that transitions from reactive crash avoidance to proactive, continuous risk management. PRISM employs inverse crash-probability modeling to convert binary crash classifiers into dynamic, interpretable safety scores. Three specialized models addressing trajectory kinematics, environmental risk, and VRU interaction operate concurrently, coordinated by a reasoning layer incorporating reinforcement learning, contextual memory, and feature-level attribution. The system provides graduated safety interventions across four tiers, from silent monitoring to emergency alerts.
Unlike rule-based systems with static thresholds, PRISM dynamically adjusts safety parameters in real time. Validated across 1,296 scenarios from three naturalistic driving datasets without dataset-specific retraining, the system yielded a mean safety score of 68 out of 100, classified 77.6% of scenarios as advisory, and flagged a near-miss rate of 3.8%, with 11% of scenarios escalating to intervention or emergency response. Feature attribution consistently identified trajectory risk and VRU proximity as primary safety factors. PRISM provides a unified, interpretable framework for proactive transportation safety with emphasis on VRU risk reduction in dense urban environments.
Artificial intelligence (AI) increasingly treats scientific literature as a data source for building databases, training predictive models, and guiding discovery. Yet literature-derived datasets often assume that reported experimental values are internally consistent and directly reusable. Here, we analyze this assumption using solid electrolyte (SE) conductivity data as a representative materials-science case. By tracing values from source articles to curated datasets, we identify recurrent text-figure mismatches, ambiguous axis annotations, unit inconsistencies, and missing measurement context. These discrepancies are often numerically plausible and therefore difficult to detect through routine preprocessing, but they can propagate as structured label noise during database construction and machine-learning reuse. A cross-database example shows how ambiguous reporting can create a 100-fold conductivity error. Our analysis reframes data accuracy as an infrastructure requirement for artificial-intelligence-driven discovery and motivates traceable reporting, curation, and validation practices for reusable scientific data.
Keywords: AI for science; Data reliability; Scientific databases; Structured label noise; Literature-derived data; Materials informatics; Solid electrolytes
Qian Wang, Ying Li, Ryuhei Sato et al.· 0 citations
Telecom Service and Network Operations Centers (SNOCs) rely on large collections of cloud documents, including Standard Operating Procedures (SOPs), vendor technical manuals, incident reports, and configuration guides, to maintain uninterrupted network operations. During critical incidents, engineers must quickly retrieve accurate information, yet traditional keyword based and single stage retrieval approaches often struggle to provide precise results.
This paper presents Athena for Cloud Knowledge Base, a fully offline, multi agent Retrieval Augmented Generation (RAG) framework designed for enterprise cloud document search in Vodafone Idea's SNOC environment. The system integrates dense retrieval using E5 Large V2 embeddings, BM25 sparse retrieval, and Knowledge Graph expansion within a LangGraph based orchestration framework. Retrieved candidates are fused using Weighted CombSUM, followed by cross encoder reranking and Maximal Marginal Relevance (MMR) to obtain a diverse and relevant evidence set.
To improve answer reliability, the framework performs per chunk LLM evaluation with explicit attribution verification, assessing each MMR selected chunk independently before generating a response. Unsupported or weak evidence is discarded, and if no chunk satisfies the verification criteria, the system automatically evaluates multiple chunks together as a fallback. Experiments on a corpus of 4200 SNOC cloud documents containing 312000 indexed chunks show that the proposed approach achieves an MRR at 10 of 0.910 and an Exact Match (EM) score of 78.4 percent, outperforming single stage dense retrieval by 14.6 percentage points. The entire pipeline operates in a fully offline environment, satisfying enterprise data sovereignty requirements while delivering accurate and grounded responses for cloud document search.
Speech brain-computer interfaces (speech BCIs) translate neural activity into language, offering a path towards restoring speech for people with paralysis and, more broadly, enabling new forms of natural human-computer interaction. Despite this promise, the field lacks a common measure of progress because systems use different datasets, recording methods, types of speech, and vocabularies, so their reported scores are rarely comparable. Underlying this measurement problem are two unresolved questions: (i) what distribution of words should a speech BCI enable a user to communicate, and (ii) how much information from this distribution can a system convey. We address both by deriving open-vocabulary mutual information (OVMI), an information-theoretic quantity that measures the information conveyed by a decoder relative to a reference distribution over the words a user may wish to communicate. This allows capabilities measured under different conditions, such as distinct vocabularies, to be evaluated on a common communication scale. We show that ordinarily reported accuracy, word error rate (WER), and other metrics computed only over the words a system supports can overstate how much of a user's intended speech the system can communicate. We then use OVMI to compare existing systems, expose trade-offs between how much of the user's language a system supports and how accurately it decodes those words, show that these comparisons depend on what the user is expected to communicate, and demonstrate that selecting a vocabulary to maximise OVMI yields up to 16.3% relative improvement in accuracy across three speech domains. OVMI therefore provides the speech BCI community with a principled way to compare heterogeneous systems, improve vocabulary design, and measure progress in the field.
Dulhan Jayalath, Benjamin Ballyk, Oiwi Parker Jones· 0 citations
A new machine-learning framework aims to improve the success rate of computational protein design while moving away from results that reproduce sequences found in nature.
MIT News · Artificial Intelligence· news.mit.eduAug 24, 2026
A new method for surgically removing training examples from a model reveals that as datasets grow, the link between what a model learns and what it produces dissolves.
