Tootfinder

“A sign of the scale of #’Twitter’s bot problem is the thriving #botindustry. Bot makers from around the world advertise their services on freelancer websites.
A computer scientist in Pakistan, sells "ChatGPT Twitter bots" for $30 to $500, depending on complexity.
In an interview with the ABC …he said could make a "fully fledged" bot that could "like comments o…

Twitter is becoming a 'ghost town' of bots as AI-generated spam content floods the internet
The internet is filling up with machine-generated "zombie content" designed to game algorithms and scam humans. Experts call it the "great AI flood".

*sarc
Sounds great over there!
P.S. Twitter is NOT the “Internet” fgs….
#TwitterEnding #Twitter #Mastodon

When you set your paperclip optimizers to optimize for "engagement," this is the obvious, inevitable result.
https://www.abc.net.au/news/science/2024-02-28/twitter-x-fighting-bot-problem-as-ai-spam-fl…

Twitter is becoming a 'ghost town' of bots as AI-generated spam content floods the internet
The internet is filling up with machine-generated "zombie content" designed to game algorithms and scam humans. Experts call it the "great AI flood".

Better & Faster Large Language Models via Multi-token Prediction
Fabian Gloeckle, Badr Youbi Idrissi, Baptiste Rozi\`ere, David Lopez-Paz, Gabriel Synnaeve
https://arxiv.org/abs/2404.19737 https://arxiv.org/pdf/2404.19737
arXiv:2404.19737v1 Announce Type: new
Abstract: Large language models such as GPT and Llama are trained with a next-token prediction loss. In this work, we suggest that training language models to predict multiple future tokens at once results in higher sample efficiency. More specifically, at each position in the training corpus, we ask the model to predict the following n tokens using n independent output heads, operating on top of a shared model trunk. Considering multi-token prediction as an auxiliary training task, we measure improved downstream capabilities with no overhead in training time for both code and natural language models. The method is increasingly useful for larger model sizes, and keeps its appeal when training for multiple epochs. Gains are especially pronounced on generative benchmarks like coding, where our models consistently outperform strong baselines by several percentage points. Our 13B parameter models solves 12 % more problems on HumanEval and 17 % more on MBPP than comparable next-token models. Experiments on small algorithmic tasks demonstrate that multi-token prediction is favorable for the development of induction heads and algorithmic reasoning capabilities. As an additional benefit, models trained with 4-token prediction are up to 3 times faster at inference, even with large batch sizes.

Private graph colouring with limited defectiveness
Aleksander B. G. Christiansen, Eva Rotenberg, Teresa Anna Steiner, Juliette Vlieghe
https://arxiv.org/abs/2404.18692 https://arxiv.org/pdf/2404.18692
arXiv:2404.18692v1 Announce Type: new
Abstract: Differential privacy is the gold standard in the problem of privacy preserving data analysis, which is crucial in a wide range of disciplines. Vertex colouring is one of the most fundamental questions about a graph. In this paper, we study the vertex colouring problem in the differentially private setting.
To be edge-differentially private, a colouring algorithm needs to be defective: a colouring is d-defective if a vertex can share a colour with at most d of its neighbours. Without defectiveness, the only differentially private colouring algorithm needs to assign n different colours to the n different vertices. We show the following lower bound for the defectiveness: a differentially private c-edge colouring algorithm of a graph of maximum degree {\Delta} > 0 has defectiveness at least d = {\Omega} (log n / (log c log {\Delta})).
We also present an {\epsilon}-differentially private algorithm to {\Theta} ( {\Delta} / log n 1 / {\epsilon})-colour a graph with defectiveness at most {\Theta}(log n).

No entiendo a la gente que sigue ahí "porque tiene mucha difusión" https://www.abc.net.au/news/science/2024-02-28/twitter-x-fighting-bot-problem-as-ai-spam-floods-the-internet/103498070

How Allegro reduced latency outliers by 82% by switching to #XFS:
https://blog.allegro.tech/2024/03/kafka-performance-analysis.html
"'Using a com…

Unlocking Kafka's Potential: Tackling Tail Latency with eBPF
At Allegro, we use Kafka as a backbone for asynchronous communication between microservices. With up to 300k messages published and 1M messages consumed every second, it is a key part of our infrastructure. A few months ago, in our main Kafka cluster, we noticed the following discrepancy: while median response times for produce requests were in single-digit milliseconds, the tail latency was much worse. Namely, the p99 latency was up to 1 second, and the p999 latency was up to 3 seconds. This w…

This https://arxiv.org/abs/2303.00614 has been replaced.
initial toot: https://mastoxiv.page/@arXiv_csNE_…

A Hybrid Genetic Algorithm with Type-Aware Chromosomes for Traveling Salesman Problems with Drone
There are emerging transportation problems known as the Traveling Salesman Problem with Drone (TSPD) and the Flying Sidekick Traveling Salesman Problem (FSTSP) that involve using a drone in conjunction with a truck for package delivery. This study presents a hybrid genetic algorithm for solving TSPD and FSTSP by incorporating local search and dynamic programming. Similar algorithms exist in the literature. Our algorithm, however, considers more sophisticated chromosomes and less computationally…

A faster algorithm for the Fr\'echet distance in 1D for the imbalanced case
Lotte Blank, Anne Driemel
https://arxiv.org/abs/2404.18738 https://<…

A faster algorithm for the Fréchet distance in 1D for the imbalanced case
The fine-grained complexity of computing the Fréchet distance has been a topic of much recent work, starting with the quadratic SETH-based conditional lower bound by Bringmann from 2014. Subsequent work established largely the same complexity lower bounds for the Fréchet distance in 1D. However, the imbalanced case, which was shown by Bringmann to be tight in dimensions $d\geq 2$, was still left open. Filling in this gap, we show that a faster algorithm for the Fréchet distance in the imbala…

@… @… definitely feeling this! Our ability to grow communities has seen incredible fragmentation in the last year and it’s resulting in stuff being seen a lot less. (Not to mention the rise of algorithmically g…