Tech Meridian
LIVE FEED
ENTITIES
EN RU

AI INDUSTRY INTELLIGENCE

WHAT MATTERS.
AS IT HAPPENS.

One event, every source. Classified, ranked and summarized in real time.

ARCHIVE DATE
346 RESULTS · PAGE 12 OF 12
RESEARCH1 SOURCE · Google Research

Google Research presents 'Talk like a Graph' and GraphQA benchmark for encoding graphs to LLMs

Researchers at Google (Bahare Fatemi and Bryan Perozzi) propose methods to translate graph-structured data into text that large language models can reason over, and introduce GraphQA, a benchmark of graph reasoning tasks and graph generators. They report that LLM performance varies with encoding method, task type, and graph structure, and that choosing the right encoding can improve graph-task performance by up to about 60%.

Why it matters: Understanding how to encode graphs for LLMs matters because many real-world data sources are graph-structured, so better encodings can enable LLMs to reason about networks, knowledge graphs and other structured data.

OPEN EVENT →
7.0IMPORTANCE
RESEARCH1 SOURCE · Google Research

Chain-of-Table: Iteratively evolving tables as a reasoning chain for improved table understanding

Researchers (Zilong Wang and Chen-Yu Lee of the Cloud AI Team) propose Chain-of-Table, a framework that trains LLMs via in-context learning to iteratively generate table operations and update intermediate tables as an explicit reasoning chain; this transforms complex tables into simpler, question-aligned views and reportedly achieves new state-of-the-art results on WikiTQ, TabFact, and FeTaQA benchmarks.

Why it matters: Chain-of-Table provides a structured, interpretable way for LLMs to reason over tabular data and reportedly improves accuracy on multiple established table-understanding benchmarks.

OPEN EVENT →
7.0IMPORTANCE
MODELS1 SOURCE · Google Research

Google releases Derm Foundation and Path Foundation embedding tools for dermatology and pathology research

Google Health / Google Research announced two domain-specific embedding tools—Derm Foundation for dermatology images and Path Foundation for digital pathology (H&E whole-slide) images—made available for research use. The tools produce specialized numerical embeddings to accelerate downstream tasks (e.g., linear-probe classifiers) and build on prior CXR Foundation work and pathology-specific self-supervised learning optimizations.

Why it matters: Domain-specific embeddings can lower the data, compute, and annotation burden for developing medical-imaging ML models and may speed research on dermatology and pathology applications.

OPEN EVENT →
7.0IMPORTANCE
RESEARCH1 SOURCE · Google Research

Social learning: Google Research framework for LLMs to teach each other via natural language

Google Research authors propose a "social learning" framework in which teacher LLMs transfer knowledge to a student LLM using natural-language instructions and synthesized examples rather than sharing raw data. Evaluated with PaLM 2‑S on tasks including spam detection, grade‑school math, and text-based question answering, they find that generated examples (e.g., 8–16 per case) often preserve privacy while yielding comparable student performance for most tasks, though spam detection was a notable exception; the paper also proposes quantitative privacy measures for this setting.

Why it matters: If LLMs can effectively teach one another using natural language and synthetic examples, teams could collaboratively improve models while reducing the need to share sensitive raw data.

OPEN EVENT →
7.0IMPORTANCE
RESEARCH1 SOURCE · Google Research

Croissant 1.0: a schema.org-based metadata format for ML-ready datasets, with support from major dataset hosts and frameworks

Croissant is a new ML-oriented metadata format (v1.0) built on schema.org that standardizes description and organization of datasets without changing underlying file formats. The release includes a spec, example datasets, an open-source Python validator/consumer/generator, a visual editor, and a Responsible AI (RAI) vocabulary extension, and is being adopted by Kaggle, Hugging Face, OpenML, indexed by Dataset Search, and made loadable in major frameworks via TensorFlow Datasets.

Why it matters: A shared ML-focused metadata format can reduce the "data development burden" by improving dataset discoverability, tooling, reuse, and tracking of Responsible AI attributes across repositories and frameworks.

OPEN EVENT →
7.0IMPORTANCE
RESEARCH1 SOURCE · Google Research

Google Quantum AI hosts booth and 50+ presentations at APS March Meeting 2024

Google’s Quantum AI team announced a prominent presence at the American Physical Society March Meeting 2024 in Minneapolis, including a booth, over 50 talks, and presentations of tools and projects such as the Crumble visualization prototype, the Qualtran resource-estimation library, and an XPRIZE-related Q&A. The schedule covers topics across quantum error correction, calibration, noisy and fault-tolerant hardware, quantum algorithms, and ML applications for materials and chemistry.

Why it matters: The showcase highlights Google’s continued investment in quantum research tools, algorithms and ML-related quantum applications, signaling potential influence on research directions and collaboration opportunities in the field.

OPEN EVENT →
6.0IMPORTANCE
MODELS1 SOURCE · Google Research

VideoPrism — Google Research's large foundational visual encoder for diverse video understanding

Google Research introduces VideoPrism, a video foundation model (ViFM) that produces frozen video representations for a wide range of tasks (classification, localization, retrieval, captioning, QA). It is pre-trained on a hybrid corpus of 36 million high-quality video-text pairs plus 582 million additional clips with noisy or machine-generated text and uses a two-stage training scheme (video-text contrastive learning followed by masked video modeling).

Why it matters: A single, widely pre-trained frozen video encoder that reportedly achieves state-of-the-art results could simplify and standardize many downstream video understanding tasks and research efforts.

OPEN EVENT →
8.0IMPORTANCE
MODELS1 SOURCE · Google Research

Google describes private on-device LM training for Gboard using federated learning and differential privacy

Google researchers report that all next-word-prediction neural LMs in Gboard are now trained with federated learning and formal differential privacy guarantees, with over 30 on-device models across 7+ languages and 15+ countries. The deployed models satisfy (ε,δ)-DP with δ=1e-10 and ε ranging from 0.994 to 13.69; the post describes practical techniques used (DP-FTRL, secure aggregation, adaptive clipping, pretraining on C4) and states this is, to their knowledge, the largest production deployment of user-level DP and the first announced ε<1 for models trained directly on user data.

Why it matters: This matters because it demonstrates production-scale deployment of federated learning with formal user-level differential privacy (including ε<1) for on-device language models, advancing practical privacy protections for user data.

OPEN EVENT →
8.0IMPORTANCE
RESEARCH1 SOURCE · Google Research

Google Research: instance-conditional timescales of decay to reweight training data under concept drift

Google Research proposes learning an auxiliary, instance-conditional weighting model that assigns importance scores to training examples as a function of their content and age, combining multiple fixed timescales of decay and meta-learning the assignment alongside the primary model. The method aims to blend benefits of offline and continual learning and yields up to ~15% relative accuracy gains on a large nonstationary photo benchmark (~39M images over 10 years) and improvements across other nonstationary learning benchmarks.

Why it matters: This approach offers a practical, learned way to reduce model degradation under slow concept drift by selectively downweighting or preserving past examples, addressing limitations of pure offline or continual strategies in real-world nonstationary data.

OPEN EVENT →
7.0IMPORTANCE
RESEARCH1 SOURCE · Google Research

Google Research releases DP-Auditorium, a Python library for black-box auditing of differential privacy

Google Research published DP-Auditorium, an open-source Python library that audits differential privacy guarantees using only black-box access to mechanisms. The library provides property testers and dataset finders (including novel HockeyStick, Rényi, and MMD testers) and demonstrates the ability to detect DP guarantee violations across multiple DP definitions (pure, approximate, Rényi).

Why it matters: Practical black-box auditing tools like DP-Auditorium help detect faulty implementations or proofs of DP mechanisms, increasing trust in privacy-preserving systems used across research and industry.

OPEN EVENT →
7.0IMPORTANCE
CODING1 SOURCE · Google Research

Google releases TensorFlow GNN 1.0 for production-scale graph neural networks

Google announces TensorFlow GNN 1.0 (TF-GNN), a production-tested library for building and training graph neural networks in TensorFlow with first-class GraphTensor support, Keras integration, and flexible dynamic or batch subgraph sampling—scalable from interactive Colab use to distributed sampling via Apache Beam for datasets of up to hundreds of millions of nodes and billions of edges.

Why it matters: TF-GNN makes heterogeneous graphs a first-class citizen in TensorFlow and provides scalable sampling and training tools, lowering the barrier to deploying GNNs on large real-world graphs.

OPEN EVENT →
7.0IMPORTANCE
MODELS1 SOURCE · Google Research

Google Research presents TimesFM, a 200M-parameter decoder-only foundation model for time-series forecasting

Google Research researchers Rajat Sen and Yichen Zhou introduce TimesFM, a decoder-only transformer foundation model pre-trained on a corpus of about 100 billion real-world time-points. At 200 million parameters, TimesFM treats contiguous time-point patches as tokens and achieves strong zero-shot forecasting on diverse unseen datasets—approaching state-of-the-art supervised methods—with plans to make the model available via Google Cloud Vertex AI later this year.

Why it matters: A pretrained time-series foundation model that delivers competitive zero-shot forecasts could reduce the need for dataset-specific retraining and speed deployment of forecasting in domains like retail and finance.

OPEN EVENT →
7.0IMPORTANCE
RESEARCH1 SOURCE · Google Research

Google Research proposes 'early readouts' and a 'feature sieve' to mitigate spurious features and simplicity bias

Google Research authors Rishabh Tiwari and Pradeep Shenoy describe two interventions—using predictions from early intermediate layers ('early readouts') and inducing 'feature forgetting' via a 'feature sieve'—to detect and reduce reliance on spurious features and the simplicity bias in deep networks. They apply the early-readout signal to reweight teacher contributions during distillation and use a feature-sieving intervention to encourage models to find more predictive features, reporting improved worst-group accuracy across benchmarks (Waterbirds, CelebA, CivilComments, MNLI) and better generalization to unseen domains compared with prior methods.

Why it matters: This matters because the methods offer automated diagnostics and practical interventions to reduce spurious-feature reliance and simplicity bias, improving worst‑group robustness and cross‑domain generalization in common benchmark tasks.

OPEN EVENT →
7.0IMPORTANCE
MODELS1 SOURCE · Google Research

MobileDiffusion: subsecond on-device text-to-image generation with a 520M-parameter latent diffusion model

MobileDiffusion is an efficient latent text-to-image diffusion model designed for mobile devices that combines a compact CLIP-ViT/L14 text encoder with a streamlined diffusion UNet and image decoder. By adopting DiffusionGAN for one-step sampling and focusing on architectural efficiency, the 520M-parameter model can produce a 512×512 image in about half a second on tested premium iOS and Android devices.

Why it matters: Faster, smaller on-device text-to-image models enable real-time mobile user experiences and improved privacy by avoiding server-side inference, making deployment of high-quality generative models on phones more feasible.

OPEN EVENT →
8.0IMPORTANCE
RESEARCH1 SOURCE · Google Research

Google Research outlines software optimizations for mixed-input matmul on NVIDIA Ampere (CUTLASS)

A Google Research blog post by Manish Gupta presents software techniques to implement mixed-input matrix multiplication (e.g., F16 inputs × U8 weights) on NVIDIA Ampere Tensor Cores by handling data-type conversion and layout conformance; the methods are released in the open-source NVIDIA/CUTLASS repository and are reported to add minimal software overhead while approaching hardware peak performance. The work targets memory- and compute-heavy LLM workloads by enabling weight-only quantization patterns that reduce model memory footprint.

Why it matters: Efficient software mapping of mixed-input matmul enables reduced-memory 8-bit weight formats for large LLMs on existing GPU hardware, improving deployment efficiency without hardware changes.

OPEN EVENT →
7.0IMPORTANCE
MODELS1 SOURCE · Google Research

Exphormer: using expander graphs to scale graph transformers

Exphormer is a sparse-attention framework for graph transformers proposed by Google Research (presented at ICML 2023) that replaces the dense full-attention interaction graph with a sparse d-regular expander graph (using a randomized Friedman near-Ramanujan construction). The approach combines expander edges with the input graph and virtual nodes to achieve linear edges and reports strong empirical results across diverse graph datasets; an implementation is available on GitHub.

Why it matters: Reducing the quadratic cost of full attention on graphs by using expander-based sparse interaction graphs makes transformer architectures practical for much larger graph-structured datasets.

OPEN EVENT →
7.0IMPORTANCE
EVENT
LOADING EVENT