Concept Bottleneck Models (CBMs) provide interpretable prediction by introducing an intermediate Concept Bottleneck Layer (CBL), which encodes human-understandable concepts to explain models' decision.

As a leading online platform with a vast global audience, YouTube's extensive reach also makes it susceptible to hosting harmful content, including disinformation and conspiracy theories. This study explores the use of open-weight Large Language Models (LLMs), both text-only and multimodal, for identifying conspiracy theory videos shared on YouTube. Leveraging a labeled dataset of thousands of videos, we evaluate a variety of LLMs in a zero-shot setting and compare their performance to a fine-tuned RoBERTa baseline. Results show that text-based LLMs achieve high recall but lower precision, leading to increased false positives. Multimodal models lag behind their text-only counterparts, indicating limited benefits from visual data integration. To assess real-world applicability, we evaluate the most accurate models on an unlabeled dataset, finding that RoBERTa achieves performance close to LLMs with a larger number of parameters. Our work highlights the strengths and limitations of current LLM-based approaches for online harmful content detection, emphasizing the need for more precise and robust systems.

In this work, we investigate how autonomous agents, organized into tribes, learn to use communication signals to coordinate their activities and enhance their collective efficiency. Using the NEC-DAC (Neurally Encoded Culture - Distributed Autonomous Communicators) system, where each agent is equipped with its own neural network for decision-making, we demonstrate how these agents develop a shared behavioral system -- akin to a culture -- through learning and signalling. Our research focuses on the self-organization of culture within these tribes of agents and how varying communication strategies impact their fitness and cooperation. By analyzing different social structures, such as authority hierarchies, we show that the culture of cooperation significantly influences the tribe's performance. Furthermore, we explore how signals not only facilitate the emergence of culture but also enable its transmission across generations of agents. Additionally, we examine the benefits of coordinating behavior and signaling within individual agents' neural networks.

Concept Bottleneck Models (CBMs) provide interpretable prediction by introducing an intermediate Concept Bottleneck Layer (CBL), which encodes human-understandable concepts to explain models' decision. Recent works proposed to utilize Large Language Models (LLMs) and pre-trained Vision-Language Models (VLMs) to automate the training of CBMs, making it more scalable and automated. However, existing approaches still fall short in two aspects: First, the concepts predicted by CBL often mismatch the input image, raising doubts about the faithfulness of interpretation. Second, it has been shown that concept values encode unintended information: even a set of random concepts could achieve comparable test accuracy to state-of-the-art CBMs. To address these critical limitations, in this work, we propose a novel framework called Vision-Language-Guided Concept Bottleneck Model (VLG-CBM) to enable faithful interpretability with the benefits of boosted performance. Our method leverages off-the-shelf open-domain grounded object detectors to provide visually grounded concept annotation, which largely enhances the faithfulness of concept prediction while further improving the model performance. In addition, we propose a new metric called Number of Effective Concepts (NEC) to control the information leakage and provide better interpretability. Extensive evaluations across five standard benchmarks show that our method, VLG-CBM, outperforms existing methods by at least 4.27% and up to 51.09% on accuracy at NEC=5, and by at least 0.45% and up to 29.78% on average accuracy across different NECs, while preserves both faithfulness and interpretability of the learned concepts as demonstrated in extensive experiments.

As repositories of large scale data in earth observation (EO) have grown, so have transfer and storage costs for model training and inference, expending significant resources. We introduce Neural Embedding Compression (NEC), based on the transfer of compressed embeddings to data consumers instead of raw data. We adapt foundation models (FM) through learned neural compression to generate multi-task embeddings while navigating the tradeoff between compression rate and embedding utility. We update only a small fraction of the FM parameters (10%) for a short training period (1% of the iterations of pre-training). We evaluate NEC on two EO tasks: scene classification and semantic segmentation. Compared with applying traditional compression to the raw data, NEC achieves similar accuracy with a 75% to 90% reduction in data. Even at 99.7% compression, performance drops by only 5% on the scene classification task. Overall, NEC is a data-efficient yet performant approach for multi-task EO modelling.

This paper compares two deep reinforcement learning approaches for cyber security in software defined networking. Neural Episodic Control to Deep Q-Network has been implemented and compared with that of Double Deep Q-Networks. The two algorithms are implemented in a format similar to that of a zero-sum game. A two-tailed T-test analysis is done on the two game results containing the amount of turns taken for the defender to win. Another comparison is done on the game scores of the agents in the respective games. The analysis is done to determine which algorithm is the best in game performer and whether there is a significant difference between them, demonstrating if one would have greater preference over the other. It was found that there is no significant statistical difference between the two approaches.

They're old ideas, but machine learning and AI are now the communications industry's hot buzzwords, judging from last week's Mobile World Congress. "For mobile operators around the world, this is no longer an experiment," said Patrick Ostiguy, CEO of Accedian, during an MWC panel discussion on machine learning. Machine learning, which involves training a computer by feeding it examples and counterexamples, has been around for decades. The post office's optical mail-sorting machines are one example. True AI and deep learning, which strive to teach a brain how to teach itself, are also long-standing disciplines.

Fox Business Flash top headlines are here. Check out what's clicking on FoxBusiness.com. TOKYO - Japanese companies are ramping up the use of artificial intelligence and other advanced technology to reduce waste and cut costs in the pandemic, and looking to score some sustainability points along the way. Disposing of Japan's more than 6 million tonnes in food waste costs the world's No.3 economy some 2 trillion yen ($19 billion) a year, government data shows. With the highest food waste per capita in Asia, the Japanese government has enacted a new law to halve such costs from 2000 levels by 2030, pushing companies to find solutions.

Japan's NEC Corp. has launched a facial recognition system that identifies people even when they are wearing masks, adapting to a new normal in which face coverings have become a key form of protection against the spread of the coronavirus. The technology firm had already been working on a system to meet the needs of allergy sufferers who wear masks when the COVID-19 pandemic prompted it to accelerate development. "Needs grew even more due to the coronavirus situation as the state of emergency (last year) was continuing for a long time, and so we've now introduced this technology to the market," Shinya Takashima, assistant manager of NEC's digital platform division, said. The system determines when a person is wearing a mask and hones in on the parts that are not covered up, such as the eyes and surrounding areas, to verify the subject's identity. Users register a photo of their face in advance.