technocracy

There is a sensation that your children will never fully understand. You are twelve years old. It is a Sunday afternoon in the dead of summer. The heat is unreasonable. There is nothing on television. Your friends are unavailable. Your parents are busy. You have already read the same three books twice. You are bored. Profoundly, magnificently, almost cosmically bored. So you lie on the floor and stare at the ceiling. And somewhere in that vast, uncomfortable emptiness […]

Ver mais

The gaming industry has a favorite excuse. When players leave, studios blame short attention spans. They blame crowded markets. They blame modern habits, social media, burnout, or the idea that players simply do not stay loyal anymore. That explanation sounds clever, but it protects the industry from asking a harder question. What if players are not quitting because they care less? What if they are quitting because too many games have become frustrating, bloated, manipulative, and exhausting to […]

Ver mais

The rapid rise in popularity of AI startup companies has caused a massive upheaval in the way we think about technology on a global scale. Due to significant technological development in areas such as machine learning (ML), data analytics, and cloud computing, the AI Startup ecosystem is now home to many successful innovators, with substantial investment and talent flocking to this space. Companies are solving extremely difficult business challenges and simultaneously changing the way businesses operate in various […]

Ver mais

arXiv:2601.19005v1 Announce Type: cross Abstract: With the advancement of machine learning and artificial intelligence technologies, recommender systems have been increasingly used across a vast variety of platforms to efficiently and effectively match users with items. As application contexts become more diverse and complex, there is a growing need for more sophisticated recommendation techniques. One example is the composite item (for example, fashion outfit) recommendation where multiple levels of user preference information might be available and relevant. In this […]

Ver mais

Federated learning (FL) enables distributed clients to collaboratively train a global model using local private data. Nevertheless, recent studies show that conventional FL algorithms still exhibit deficiencies in privacy protection, and the server lacks a reliable and stable aggregation rule for updating the global model. This situation creates opportunities for adversaries: on the one hand, they may eavesdrop on uploaded gradients or model parameters, potentially leaking benign clients’ private data; on the other hand, they may compromise clients […]

Ver mais

Active learning reduces labeling costs by selecting samples that maximize information gain. A dominant framework, Query-by-Committee (QBC), typically relies on perturbation-based diversity by inducing model disagreement through random feature subsetting or data blinding. While this approximates one notion of epistemic uncertainty, it sacrifices direct characterization of the plausible hypothesis space. We propose the complementary approach: Rashomon Ensembled Active Learning (REAL) which constructs a committee by exhaustively enumerating the Rashomon Set of all near-optimal models. To address functional redundancy […]

Ver mais

arXiv:2405.17490v4 Announce Type: replace-cross Abstract: Influence functions serve as crucial tools for assessing sample influence in model interpretation, subset training set selection, noisy label detection, and more. By employing the first-order Taylor extension, influence functions can estimate sample influence without the need for expensive model retraining. However, applying influence functions directly to deep models presents challenges, primarily due to the non-convex nature of the loss function and the large size of model parameters. This difficulty not only makes […]

Ver mais

arXiv:2604.09731v1 Announce Type: new Abstract: Tree-based speculative decoding accelerates autoregressive generation by verifying a branching tree of draft tokens in a single target-model forward pass. However, existing methods prioritize maximizing token-level likelihood or the number of accepted tokens while ignoring a critical “efficiency paradox”: the computational overhead of drafting and verifying big trees can grow super-linearly, particularly at scale. This often leads to negative wall-clock speedup when batch sizes increase or hardware saturation limits are reached. To address […]

Ver mais