This paper advances NLP research for the low-resource Uzbek language by evaluating two previously untested monolingual Uzbek BERT models on the part-of-speech (POS) tagging task and introducing the first publicly available UPOS-tagged benchmark dataset for Uzbek. Our fine-tuned models achieve 91% average accuracy, outperforming the baseline multi-lingual BERT as well as the rule-based tagger. Notably, these models capture intermediate POS changes through affixes and demonstrate context sensitivity, unlike existing rule-based taggers.

Decoding from the output distributions of large language models to produce high-quality text is a complex challenge in language modeling. Various approaches, such as beam search, sampling with temperature, $k-$sampling, nucleus $p-$sampling, typical decoding, contrastive decoding, and contrastive search, have been proposed to address this problem, aiming to improve coherence, diversity, as well as resemblance to human-generated text. In this study, we introduce adaptive contrastive search, a novel decoding strategy extending contrastive search by incorporating an adaptive degeneration penalty, guided by the estimated uncertainty of the model at each generation step. This strategy is designed to enhance both the creativity and diversity of the language modeling process while at the same time producing coherent and high-quality generated text output. Our findings indicate performance enhancement in both aspects, across different model architectures and datasets, underscoring the effectiveness of our method in text generation tasks. Our code base, datasets, and models are publicly available.

OpenAI has released the Chat Generative Pre-trained Transformer (ChatGPT) and revolutionized the approach in artificial intelligence to human-model interaction. Several publications on ChatGPT evaluation test its effectiveness on well-known natural language processing (NLP) tasks. However, the existing studies are mostly non-automated and tested on a very limited scale. In this work, we examined ChatGPT's capabilities on 25 diverse analytical NLP tasks, most of them subjective even to humans, such as sentiment analysis, emotion recognition, offensiveness, and stance detection. In contrast, the other tasks require more objective reasoning like word sense disambiguation, linguistic acceptability, and question answering. We also evaluated GPT-4 model on five selected subsets of NLP tasks. We automated ChatGPT and GPT-4 prompting process and analyzed more than 49k responses. Our comparison of its results with available State-of-the-Art (SOTA) solutions showed that the average loss in quality of the ChatGPT model was about 25% for zero-shot and few-shot evaluation. For GPT-4 model, a loss for semantic tasks is significantly lower than for ChatGPT. We showed that the more difficult the task (lower SOTA performance), the higher the ChatGPT loss. It especially refers to pragmatic NLP problems like emotion recognition. We also tested the ability to personalize ChatGPT responses for selected subjective tasks via Random Contextual Few-Shot Personalization, and we obtained significantly better user-based predictions. Additional qualitative analysis revealed a ChatGPT bias, most likely due to the rules imposed on human trainers by OpenAI. Our results provide the basis for a fundamental discussion of whether the high quality of recent predictive NLP models can indicate a tool's usefulness to society and how the learning and validation procedures for such systems should be established.

Despite the seeming success of contemporary grounded text generation systems, they often tend to generate factually inconsistent text with respect to their input. This phenomenon is emphasized in tasks like summarization, in which the generated summaries should be corroborated by their source article. In this work, we leverage recent progress on textual entailment models to directly address this problem for abstractive summarization systems. We use reinforcement learning with reference-free, textual entailment rewards to optimize for factual consistency and explore the ensuing trade-offs, as improved consistency may come at the cost of less informative or more extractive summaries. Our results, according to both automatic metrics and human evaluation, show that our method considerably improves the faithfulness, salience, and conciseness of the generated summaries.

The protection of non-combatants in times of (fully) autonomous warfare raises the question of the timeliness of the international protective emblem. Incidents in the recent past indicate that it is becoming necessary to transfer the protective emblem to other dimensions of transmission and representation. (Fully) Autonomous weapon systems are often launched from a great distance to the aiming point and there may be no possibility for the operators to notice protective emblems at the point of impact. In this case, the weapon system would have to detect such protective emblems and, if necessary, disintegrate autonomously or request an abort via human-in-the-loop. In our paper, we suggest ways in which a cross-frequency protective emblem can be designed. On the one hand, the technical deployment, e.g. in the form of RADAR beacons, is considered, as well as the interpretation by methods of machine learning. With regard to the technical deployment, possibilities are considered to address different sensors and to send signals out as resiliently as possible. When considering different signals, approaches are considered as to how software can recognise the protective emblems under the influence of various boundary conditions and react to them accordingly. In particular, a distinction is made here between the recognition of actively emitted signals and passive protective signals, e.g. the recognition of wounded or surrendering persons via drone-based electro-optical and thermal cameras. Finally, methods of distribution are considered, including encryption and authentication of the received signal, and ethical aspects of possible misuse are examined.

The U.S. military has launched two drone strikes against Taliban positions in northern Afghanistan, a U.S. defense official told Fox News on Friday. The strikes came hours before Afghan President Ashraf Ghani is set to meet President Biden at the White House Friday afternoon. An unknown number of Taliban fighters were killed in Baghlan and Kunduz provinces, where the strikes took place. Taliban fighters react to a speech by their senior leader in the Shindand district of Herat province, Afghanistan, in 2016. Three more districts fell to the Taliban overnight.

Indian and Pakistani troops fired at each other in disputed Kashmir on Monday, as Indian troops searched an army camp elsewhere in the region where suspected militants killed an Indian paramilitary soldier. Indian army Lt. Col. Manish Mehta said Pakistani troops fired without provocation using small arms and mortar shells in the Poonch sector of the Line of Control separating the Indian- and Pakistani-controlled parts of Kashmir. Pakistan's army said in a statement that its troops were responding to unprovoked firing by Indian soldiers. Both sides said the exchange of fire was continuing. In Islamabad, Prime Minister Nawaz Sharif met with the leaders of all Pakistani political parties to discuss the ongoing clashes.

WASHINGTON/PESHAWAR, PAKISTAN – The leader of the Islamic State group's branch in Afghanistan and Pakistan was killed in a U.S. drone strike on July 26, a Pentagon spokesman said on Friday after the Afghan ambassador to Pakistan announced the news to Reuters. The death of Hafiz Saeed Khan is a blow to efforts by the Islamic State -- also known as ISIS or Daesh -- to expand from its heartlands in Syria and Iraq into Afghanistan and Pakistan, which already are crowded with jihadi movements, including the Taliban and al-Qaida. It is the second U.S. killing of a prominent militant in the region in months. In May, a U.S. drone killed Afghan Taliban leader Mullah Akhtar Mansour in a strike in Pakistan. Despite that, Afghanistan's 15-year-old war grinds on with no clear victory in sight.

It's a weird feeling, cruising around Silicon Valley in a car driven by no one. I am in the back seat of one of Google's self-driving cars – a converted Lexus SUV with lasers, radar and low-res cameras strapped to the roof and fenders – as it manoeuvres the streets of Mountain View, California, not far from Google's headquarters. I grew up about eight kilometres from here and remember riding around on these same streets on a Schwinn Sting-Ray. Now, I am riding an algorithm, you might say – a mathematical equation, which, written as computer code, controls the Lexus. The car does not feel dangerous, nor does it feel like it is being driven by a human. It rolls to a full stop at stop signs, veers too far away from a delivery van, taps the brakes for no apparent reason as we pass a line of parked cars. I wonder if the flaw is in me, not the car: Is it reacting to something I can't see? The car is capable of detecting the motion of a cat, or a car crossing the street hundreds of metres away in any direction, day or night (snow and fog can be another matter). "It sees much better than a human being," Dmitri Dolgov, the lead software engineer for Google's self-driving-car project, says proudly. He is sitting behind the wheel, his hands on his lap. As we stop at the intersection, waiting for a left turn, I glance over at a laptop in the passenger seat that provides a real-time look at how the car interprets its surroundings. On it, I see a gridlike world of colourful objects – cars, trucks, bicyclists, pedestrians – drifting by in a video-game-like tableau. Each sensor offers a different view – the lasers provide three-dimensional depth, the cameras identify road signs, turn signals, colours and lights. The computer in the back processes all this information in real time, gauging the speed of oncoming traffic, making a judgment about when it is OK to make a left turn. Waiting for the car to make that decision is a spooky moment. I am betting my life that one of the coders who worked on the algorithm for when it's safe to make a left-hand turn in traffic had not had a fight with his girlfriend (or boyfriend) the night before and screwed up the code.