Topic
research
Technology CMFRI Launches DeepDATA Mobile App for Digital Documentation of Deep-Sea Fish Biodiversity
The ICAR-Central Marine Fisheries Research Institute (ICAR-CMFRI) has launched DeepDATA, a mobile application for digitally documenting deep-sea marine biodiversity. The app allows users to browse verified information on deep-sea fish species and contribute field observations, which are then validated by CMFRI experts before being added to a national digital repository.
Commodities IISR and Spices Board Partner to Boost Spice Startups and Exports via Incubation
The ICAR-Indian Institute of Spices Research (IISR) has partnered with the Spices Board through its Agventure Technology Business Incubator (TBI) to support spice startups and export-oriented enterprises. The MoU establishes Agventure TBI as an incubation centre under the ASPIRE scheme, offering research commercialization, technology transfer, and market access.
Technology Hyderabad Researchers Develop AI-Powered Plant Leaf Disease Detection System with 96% Accuracy
A team led by Vijaya Saraswathi at VNR Vignana Jyothi Institute of Engineering and Technology in Hyderabad has patented an AI-powered leaf disease detection system that uses a convolutional neural network trained on over 20,000 images to identify diseases in tomato, potato, and pepper crops with 96% accuracy. The system also recommends pesticides and is planned for mobile app deployment.
Technology The Chatbot That Foretold Why People Share Secrets With ChatGPT
A new book, 'Inventing ELIZA', recovers the source code of the 1960s chatbot from MIT Archives. The 'ELIZA effect' shows how people attribute empathy to computers, with profound implications for modern AI trust and enterprise deployment.
Scientists Use AI and Quantum Computing to Generate New Peptides in Spare Time
Researchers at the Technical University of Denmark used a hybrid AI-quantum computing system to generate novel peptides, achieving better results than classical models especially with limited data. The work, done on weekends with leftover funds, could accelerate personalized immunotherapies and vaccines.
AI4SE and SE4AI Exploration: A Decade Review Identifies Research Gaps for Enterprise Tech Leaders
A new paper on arXiv traces a decade of progress in AI for systems engineering (AI4SE) and SE for AI (SE4AI). Using human and AI raters, it reviewed 1,712 INCOSE INSIGHT articles and 889 SERC publications, identifying five critical research gaps. The authors also launched an interactive web application for practitioners.
SoftSkill: Compressing AI Agent Skills into Compact Latent Controls Boosts Accuracy Over Traditional Prompting
Researchers propose SoftSkill, a method that compresses natural-language agent skills into compact continuous vectors, improving accuracy on benchmarks like LiveMath by 42.1 points over no-skill prompting. The approach uses a frozen backbone and a trainable soft delta, offering a more efficient alternative to traditional Markdown skill files.
New Research Shows Pretraining Data Composition Can Engineer Neural Scaling Laws for Particle Physics
A new arXiv paper demonstrates that neural scaling laws in particle physics can be engineered by adjusting pretraining data composition. The study shows that including more diverse and task-aligned synthetic data can shift scaling behavior to require more data rather than larger models, offering insights for efficient AI training.
PrefSQA Introduces Pairwise Preference Prediction for Speech Quality Assessment
A research paper proposes PrefSQA, a pairwise preference prediction method for speech quality assessment that reduces rater variability compared to traditional mean opinion scores. The method incorporates uncertainty-aware logits, an impairment attention head, and non-matching-reference comparisons. Experiments on five datasets show clear improvements over baselines, especially with high-quality preference data.
Researchers Propose Feature Selection to Improve Neural Additive Model Efficiency and Interpretability
A research paper proposes adding feature selection mechanisms to Neural Additive Models (NAM) and Neural Basis Models (NBM) to reduce computational costs and enable handling of feature interactions in high-dimensional datasets. The method updates selection weights during training, achieving better or comparable performance to state-of-the-art GAMs.
New Research Reveals How Visual Tokens Evolve Inside Vision-Language Models
A new computer vision paper from arXiv investigates how visual tokens are integrated into large language models (LLMs) under two paradigms: in-context prompting and layer-wise injection. The authors find that visual tokens enter the LLM as 'disguised visual context' lacking linguistic structure, then evolve differently depending on the integration architecture. They show that attention allocation alone is insufficient, and performance depends on the quality of visual representations at each layer.
Sequential DPO Study Reveals Non-Uniform Forgetting Across Multiple Preference Objectives
A study by Bhandari et al. on sequential Direct Preference Optimization (DPO) finds that later training objectives do not uniformly degrade earlier preferences. Using Llama-3.1-8B-Instruct, the research reveals that forgetting patterns vary from stability to positive transfer depending on objective compatibility and signal strength, offering guidance for multi-objective AI alignment in enterprises.
New Training-Free Method Enables Robots to Follow Personalized Commands Like 'Bring My Cup'
Researchers propose Visual Attentive Prompting (VAP), a training-free perceptual adapter that enables vision-language-action models to follow personalized commands by using reference images as visual prompts. VAP outperforms generic policies and token-learning baselines on simulation and real-world benchmarks.
LOKI Memory-Free Method Improves Lifelong Knowledge Editing in Language Models by 14%
Researchers introduce LOKI, a memory-free method for lifelong knowledge editing in language models. It uses dynamic layer selection via the Hilbert-Schmidt Independence Criterion and projects gradient updates onto the null-space of model weights, eliminating the need for previous knowledge access. Experiments show up to 14% improvement in average accuracy over existing approaches.
Bid Farewell to Seesaw: New Framework Boosts Long-Tail Recommendation Accuracy Without Sacrificing Diversity
A new research paper from arxiv introduces HID (Hybrid Intent-based Dual Constraint Framework), a plug-and-play solution for session-based recommendation that simultaneously improves recommendation accuracy and long-tail item performance, eliminating the traditional trade-off. The framework uses hybrid intent learning and intent constraint loss to filter session-irrelevant noise.
MENTOR: Reinforcement Learning via Flexible Teacher-Optimized Rewards for Tool-Use Distillation
A new research paper introduces MENTOR, a reinforcement learning framework that uses flexible teacher-optimized rewards to distill tool-use capabilities from large language models into small models. The approach improves out-of-domain generalization compared to supervised fine-tuning and strict reinforcement learning baselines.
TeleMorpher: New AI Framework Edits Video Motion and Location Simultaneously
Researchers have developed TeleMorpher, a one-shot framework for simultaneous motion and location editing in video. The approach leverages motion priors, segmentation, and training-free pose warping to achieve robust edits while preserving appearance. Experiments show superior performance on in-the-wild videos and the TaiChi dataset.
New AI Framework PEGE Boosts HIV Detection by 15.4% in Networked Testing
A new AI framework called Policy-Embedded Graph Expansion (PEGE), combined with Dynamics-Driven Branching (DDB), improves HIV detection by 15.4% in networked testing. Developed with WHO and University of Witwatersrand, the approach supports UN Sustainable Development Goal 3.3 by making testing more efficient on incrementally revealed disease networks.
British Space Startup Mass Balance Launches Autonomous Longevity Lab Into Orbit on SpaceX
British startup Mass Balance launched a self-run chemical laboratory into orbit on a SpaceX transporter to study disease-causing proteins under microgravity. The grapefruit-sized pod will autonomously measure cell behavior and beam data back to Earth, aiming to generate training data for AI models like AlphaFold to improve predictions for age-related diseases.
Manufacturing ICAR-IISR Signs Spice Technology Licensing Deals on 51st Foundation Day
ICAR-Indian Institute of Spices Research (IISR) marked its 51st foundation day by signing technology licensing agreements for improved spice varieties and value-added products with the State Kudumbashree Mission and Gramin Multi State Agro Cooperative Society Limited, targeting wider adoption among spice growers and food processors.
Commodities Jayant Agro Develops Short-Duration Castorseed with ICAR and University Partners
Jayant Agro-Organics Ltd (JAOL) is collaborating with ICAR's Indian Institute of Oilseed Research and Sardar Dantiwada University to develop short-duration castor varieties. Managing Director Abhay Udeshi says the initiative aims to encourage farmers to increase castorseed area by reducing crop duration. Current yields have risen from 600-700 kg/ha to 2-2.5 tonnes, and the company targets 3 tonnes nationally within 5-10 years, despite erratic monsoon and heat waves affecting this year's crop.
Technology India and Switzerland Step Up Innovation Partnership with Focus on Startups, Research
India and Switzerland are enhancing their bilateral innovation partnership, leveraging Swissnex to connect startups, researchers, and industry. Key pillars include the Indo-Swiss Joint Research Programme, healthcare collaborations, and growing interest in India's digital public infrastructure. New science and technology initiatives are expected later this year.
DRFLOW Benchmark Targets Personalized Workflow Prediction for Enterprise AI Agents
Researchers introduce DRFLOW, a benchmark for evaluating AI agents on predicting personalized workflows from heterogeneous sources. The benchmark contains 100 tasks across five domains with 1,246 workflow steps grounded in over 3,900 sources, and defines seven diagnostic metrics. A reference agent, DRFLOW-Agent, shows improvement over baselines but highlights significant remaining challenges.
Multi-Agent Reinforcement Learning Achieves Superhuman Racing with 50% Fewer Collisions
A new study demonstrates that multi-agent reinforcement learning (MARL) allows quadrotors to achieve superhuman racing performance. Agents trained via league-based self-play outperformed champion humans at over 22 m/s and cut collision rates by 50% versus single-agent baselines, suggesting a new path for safe autonomous systems in shared spaces.
Residual-Space Evolutionary Optimization via Flow-based Generative Models
A new framework called residual-space evolutionary optimization addresses the challenge of data editing with non-differentiable objectives in flow-based generative models. By operating in residual space, it separates local exploitation (self-pollination) from broader exploration (cross-pollination). The method was validated on the MorphoMNIST benchmark and crystal data, showing balanced target alignment, instance preservation, and diversity.
CADBench: A Multimodal Benchmark for AI-Assisted CAD Program Generation
CADBench is a unified benchmark for multimodal CAD program generation, containing 18,000 evaluation samples across six benchmark families, five input modalities, and six metrics. The benchmark evaluates eleven AI systems, generating over 1.4 million CAD programs, and reveals key failure modes in current approaches.
Learning What to Remember: Observability-Safe Memory Retention via Constrained Optimization for Long-Horizon Language Agents
A new research paper formulates memory retention in long-horizon language agents as a constrained stochastic optimization problem, proposing OSL-MR (Observability-Safe Learning for Memory Retention). The method combines an evidence learner with a Mixed-Score heuristic, achieving superior performance under tight budgets on benchmarks LoCoMo and LongMemEval. The work establishes a principled foundation for memory management in AI agents.
MEAL Benchmark Enables Continuous Multi-Agent RL Training on 100 Tasks in Hours Using GPU Acceleration
Researchers introduced MEAL (Multi-agent Environments for Adaptive Learning), the first benchmark for continual multi-agent reinforcement learning. Using JAX and GPU acceleration, MEAL enables training on sequences of 100 tasks in hours on a single GPU, revealing failure modes not apparent at smaller scales. This addresses the limitation of previous benchmarks that only considered 3-10 sequential tasks due to CPU constraints.
New AI Research Shows Vision-Language Models Think Better with Visual Grounding
Researchers introduce visually grounded thinking, a reasoning process that interleaves natural-language thoughts with explicit point or box groundings to image regions. The method, using a scalable synthesis pipeline and grounding-aware reinforcement learning, consistently improves performance of Gemma3-4B-IT on counting and spatial reasoning benchmarks, with the 4B model matching or surpassing the 27B variant.
DF3DV-1K: Large-Scale Dataset and Benchmark for Distractor-Free Novel View Synthesis
Researchers introduced DF3DV-1K, a large-scale real-world dataset with 1,048 scenes and 89,924 images for distractor-free novel view synthesis. The dataset spans 128 distractor types and 161 scene themes, enabling benchmarking of nine radiance field methods and 3D Gaussian Splatting. Fine-tuning a diffusion-based 2D enhancer on DF3DV-1K achieved average improvements of 0.96 dB PSNR and 0.057 LPIPS.
Latent Gaussian Splatting Achieves State-of-the-Art in 4D Panoptic Occupancy Tracking for Robots
Researchers introduce Latent Gaussian Splatting (LaGS) for 4D Panoptic Occupancy Tracking, a method that models 3D features as sparse Gaussians for continuous spatiotemporal scene understanding. It achieves state-of-the-art results on Occ3D nuScenes and Waymo datasets, addressing limitations of bounding box tracking and static occupancy estimation.
FM-Agent: New Framework Automates Formal Code Verification for Large-Scale LLM-Generated Software
FM-Agent, a new framework from researchers, automates compositional reasoning for large-scale systems using LLMs. It generates function-level specifications from caller expectations, enabling verification against natural-language intent. In evaluation, it found 522 new bugs in systems up to 143,000 lines of code within 2 days.
Vero: An Open RL Recipe for General Visual Reasoning — A Fully Open Vision-Language Model Family
A new research paper introduces Vero, a family of fully open vision-language models (VLMs) that use reinforcement learning (RL) to achieve strong general visual reasoning. The team constructed a 600K-sample dataset from 59 datasets and designed task-routed rewards. Vero variants outperformed their base models by 2.9-5.4 points on average across a 30-benchmark suite, and the best variant surpassed a stronger closed model by 3.8 points. All code, data, and models are released publicly.
SafeSpec: New Framework Boosts LLM Safety Without Sacrificing Inference Speed
Researchers propose SafeSpec, a safety-aware speculative inference framework that attaches a latent safety head to jointly evaluate semantic validity and safety in a single forward pass. On Qwen3-32B, it reduces attack success rates by 15% while preserving a 2.06x inference speedup on benign workloads, addressing the fundamental incompatibility between existing safety methods and speculative decoding.
AI-Enhanced Neural Network Auto-Tunes Quantum Dot Simulators for Majorana Mode Discovery
Researchers propose a neural network-based model that learns the landscape of quantum dot simulators and autotunes them toward Majorana modes. The deep vision-transformer network, trained on synthetic conductance maps with a physics-informed loss, can drive a quantum dot chain to a topological phase in a single update step from a broad range of initial detunings.
RTSGameBench Benchmark Tests Strategic Reasoning in Vision-Language Models
A new benchmark called RTSGameBench evaluates strategic reasoning in vision-language models (VLMs) using the real-time strategy game Beyond All Reason. The benchmark includes diagnostic mini-games, diverse matchup structures, and a self-evolving generation framework. Initial tests show state-of-the-art VLMs struggle with tighter coordination, multiagent tasks, and increased scale.
Hierarchical BART strategy achieves state-of-the-art Vietnamese multi-document summarization
A research team presents a novel hierarchical BART-based strategy for Vietnamese multi-document abstractive summarization, achieving a ROUGE2-F1 score of 0.2468 on the VLSP 2022 public test set. The approach condenses documents guided by a golden summary, producing fluent and concise outputs, and releases additional training data to the community.
ArXiv Paper Introduces KG-SoftMAP: Bayesian Network Learning from Sparse Data Using Knowledge Graph Priors
KG-SoftMAP is a new method for Bayesian network structure learning from sparse discrete data, using a weighted knowledge graph as a prior. On synthetic benchmarks, it achieved Directed-F1 scores up to 0.97 at higher observation rates, while data-only learners stayed near zero. On real educational datasets, it matched logistic regression within 0.03 F1_FAIL while providing an interpretable concept graph.
New Research Provides Conditional Diffusion Guidance Under Hard Constraints for AI
A research paper proposes a framework for conditional generation in diffusion models under hard constraints, using Doob's h-transform and martingale-based learning algorithms. The method guarantees constraint satisfaction with probability one, targeting safety-critical applications and rare-event simulation.
ACUTE Protocol Improves LLM Calibration and Trustworthiness with Activation-Based Confidence Estimates
A new research protocol, ACUTE, leverages model activations to produce better-calibrated confidence estimates for large language models. Combined with a novel metric called EURO that balances calibration and informativeness, ACUTE outperforms baselines across multiple tasks and model families, offering enterprises a path to more trustworthy AI outputs.
New AI Model Lets Robots Grasp Objects Like Humans Using RGB-D Data
Researchers introduce HUG, a flow-matching AI model that generates diverse human grasps for any object from a single RGB-D image. Trained on the 1M-HUGs egocentric dataset of 1 million frames from human grasp demonstrations, HUG outperforms state-of-the-art baselines by 23% and 34% on a challenging benchmark, enabling zero-shot grasping for multi-fingered robots.
Dual-Agent Framework Translates Natural-Language Lab Protocols Into Robotic Execution
Researchers from an unnamed institution propose a dual-agent framework that translates natural-language microplate-based biological protocols into executable robotic commands. The system uses a Parser Agent and a Heterogeneous LLM Validation Agent with a self-correction loop, evaluated across 7 parsers and 3 validators on ELISA and Bradford assays.
LedgerAgent: A New Method for Policy-Adherent Tool-Calling AI Agents in Customer Service
Researchers introduce LedgerAgent, an inference-time method that maintains observed task states in a separate ledger and checks policy constraints before tool calls, improving pass^k metrics across four customer-service domains. The approach addresses common failure modes where agents use stale or incorrect information or violate domain policies.
STAR Allocation Method Improves Text-to-Image AI Training with Spatiotemporal Rewards
A new method called SpatioTemporal Adaptive Reward (STAR) Allocation improves reinforcement learning post-training for text-to-image generation. By using text-image attention to allocate rewards to relevant latent regions, STAR enhances compositional semantic alignment, text rendering, and preference optimization without changing the external reward source. The method was validated on Stable Diffusion 3.5 Medium, achieving top scores on GenEval, OCR, and PickScore benchmarks.
The Scaffold Effect: How Prompt Framing Skews AI Evaluation in Clinical Vision-Language Models
A study on arXiv evaluating 12 open-weight vision-language models (VLMs) on clinical neuroimaging datasets found that up to 58% of apparent multimodal performance gains are due to prompt framing rather than genuine reasoning. The researchers identified a 'scaffold effect' where merely mentioning MRI availability in the task prompt accounts for 70-80% of F1 improvement, even when no imaging data is present. Expert evaluation also revealed fabrication of neuroimaging-grounded justifications, raising concerns about the reliability of VLM evaluations in clinical settings.
Study Reveals How Mixed Compliance Demonstrations Affect LLM Safety Alignment
A recent paper investigates how safety-aligned large language models interpret mixed compliance demonstrations, finding that benign demonstrations can either reduce or increase harmful compliance depending on the model. Preference optimization and demonstration ordering are critical factors.
Leveraging Non-Linearity to Overcome Data Scarcity in Intelligent Fault Diagnosis Systems
A new research paper proposes a method to design Intelligent Fault Diagnosis Systems under strong data scarcity by leveraging intrinsic non-linearities of systems. The approach uses a periodic multi-excitation level procedure and pre-trained Convolutional Neural Networks, validated on a railway pantograph structure.
RoboSSM Introduces State-Space Models for Scalable In-Context Imitation Learning in Robotics
RoboSSM is a new method for in-context imitation learning (ICIL) that replaces Transformer-based architectures with state-space models (SSMs). The approach uses Longhorn, a state-of-the-art SSM, enabling linear-time inference and strong extrapolation to longer prompts. Experiments on the LIBERO benchmark show improved generalization to unseen and long-horizon tasks compared to Transformer-based ICIL methods.
Dysarthric Speech Recognition Improved by 4.65% with F-TDNN Model and Pitch Features
A systematic study by researchers from multiple institutions investigates dysarthric speech recognition using spectral features and acoustic models. The study, published on arXiv, demonstrates that incorporating pitch features and using the Factorized Time Delay Neural Network (F-TDNN) model yields a 4.65% relative improvement in isolated word recognition and a 4.63% relative improvement in sentence recognition for dysarthric speech, compared to previous research.
AAPA: Adversarially Anchored Preference Alignment Enhances LLM Post-Training Performance
Researchers propose AAPA, a plug-in framework that adds a sentence-level adversarial anchoring signal to existing post-training objectives for large language models. Experiments on instruction-following benchmarks show consistent improvements, with staged AAPA achieving 5.77% gain on Qwen3-0.6B and 3.75% on Qwen3-4B over a strong GRPO baseline.
UniMM Framework Achieves State-of-the-Art in Multi-Agent Simulation for Autonomous Driving
Researchers introduce UniMM, a unified mixture model framework for multi-agent simulation that covers regression-based and discrete models. The framework achieves state-of-the-art performance on the WOSAC benchmark by addressing behavioral multimodality and closed-loop distributional shifts.
ParaScale: A Gauge-Invariant Approach to Scale-Calibrated Camera-Motion Transfer
Researchers present ParaScale, a plug-and-play module that calibrates camera-motion transfer between videos of vastly different scales using a gauge-invariant parallax number. It reduces parallax consistency error by more than 3x over uncalibrated methods.
Efficient and Sound Probabilistic Verification Secures AI Agents Against Policy Violations
Researchers introduce a sound and efficient framework for probabilistic verification of AI agents, addressing the need for enforcing security policies under ambiguity. The approach computes upper bounds on violation probability without independence assumptions, outperforming prior art on standard benchmarks.
New Algorithm SLARouter Cuts LLM Inference Costs by 2.2x While Guaranteeing SLA Compliance
A new research paper introduces SLARouter, an online routing algorithm that learns a cost-optimal policy from sparse user feedback while guaranteeing SLA compliance. Experiments show up to 2.2x cost reduction over existing baselines without per-benchmark tuning.
New Research Proposes Adversarial Reweighting to Calibrate Mixture-of-Experts Models Under Distribution Shift
A new study examines how mixture-of-experts (MoE) models behave under distribution shift and proposes an adversarial reweighting approach to maintain calibration. The method improves the accuracy-calibration tradeoff across model classes, prediction tasks, and distribution shifts.
Before the Labels: How Dataset Construction Biases Suicidality Detection in Clinical Text
A new paper from arXiv argues that clinical NLP datasets built from electronic health records encode specific operationalizations of suicidality, shaped by governance constraints, ICD-based cohort selection, and annotation practices. The authors demonstrate that identical labels can subsume heterogeneous clinical framings, raising concerns for AI-driven healthcare decisions.
Triangular Consistency Constraint Offers Universal Plug-and-Play Component for Optical Flow Learning
Researchers propose triangular consistency, a first-principled constraint for optical flow that is agnostic to network architecture, supervision type, and dataset. The constraint composes two flows to induce a third and enforces consistency, showing consistent improvement across supervised, unsupervised, and transfer learning with negligible computational overhead.
Lagrange: New Open-Vocabulary Sparse Framework Promises Robust Autonomous Driving in Open Worlds
A new framework called Lagrange, based on Masked Latent Fields and vision-language models, aims to enable autonomous vehicles to handle out-of-distribution scenarios and produce kinematically valid trajectories. Offline evaluations on nuScenes and CODA benchmarks show promising results for robust open-world driving.
ITNet: A Learnable Integral Transform That Unifies Convolution, Attention, and Recurrence in One Architecture
Researchers introduce ITNet, a neural architecture built on a learnable integral transform that subsumes convolution, self-attention, and recurrence as special cases. A single ITNet with a shared operator matches or exceeds specialized models on ImageNet-1K, GLUE, ModelNet40, VQA v2, and NLVR2, enabled by efficient techniques such as tiled kernel fusion and Monte Carlo integration.
RACL: Reasoning-Agent Control Layers Show Promise for Continuous Metaheuristic Learning in Vehicle Routing
A new research paper introduces Reasoning-Agent Control Layers (RACL), a method that places an AI reasoning agent above existing metaheuristic optimizers to discover and validate control rules. Tested on vehicle routing, RACL outperformed baseline policies in 21 of 21 cases, achieving average cost reductions of up to 8.3%.