Search

External pressure suppresses the ferromagnetism of localized Cr 3d electron moments in the van der Waals insulator CrBr3, which cannot be explained without considering a dramatic pressure-induced crystal or electronic structure change. We addressed this issue by conducting a parallel experimental investigation of single crystals magnetic and structural properties using magnetization, X-ray diffraction, and Raman spectroscopy measurements. Ab initio DFT calculations of electronic structure and atomistic simulations of finite-temperature magnetism supported the analysis and interpretation of experimental results. The magnetization measurements at high pressures provided the first direct experimental evidence of the pressure-induced suppression of ferromagnetism. We observed a gradual decrease of the bulk magnetic moment and Curie temperature with increasing pressure, which accelerates at pressures above 3 GPa, leading to loss of ferromagnetism at 6.5 GPa. By increasing pressure, the ambient pressure phase gradually breaks down and is accompanied by the generation of layer stacking that favor the antiferromagnetic coupling of Cr moments. As a result, the appearing antiparallel pairs of moments disturb the ferromagnetic structure and reduce the bulk magnetic moment and Curie temperature. This scenario, which is well corroborated by the results of our theoretical calculations, suggests an antiferromagnetic phase emerging with increasing pressure beyond the critical value when the new single trigonal P-3m1 phase becomes stable, characterized by the "antiferromagnetic" A-A layer stacking, becomes stable. The weak coupling between adjacent magnetic layers in van der Waals materials allows variations in layering due to sufficient external forces. We hope our comprehensive study's results can help other researchers resolve frequently appearing issues of similar origin in this class of materials.

In this work we report on a self-assembled growth of a Ge quantum dot lattice in a single 600-nm-thick Ge+Al2O3 layer during magnetron sputtering deposition of a Ge+Al2O3 mixture at an elevated substrate temperature. The self-assembly results in the formation of a well-ordered threedimensional body-centered tetragonal quantum dot lattice within the whole deposited volume. The quantum dots formed are very small in size less than 4.0 nm, have a narrow size distribution and a large packing density. The parameters of the quantum dot lattice can be tuned by changing the deposition parameters. The self-ordering of the quantum dots is explained by diffusionmediated nucleation and surface-morphology effects and simulated by a kinetic Monte Carlo model.

Multilingual large language models (LLMs) have gained prominence, but concerns arise regarding their reliability beyond English. This study addresses the gap in cross-lingual semantic evaluation by introducing a novel benchmark for cross-lingual sense disambiguation, StingrayBench. In this paper, we demonstrate using false friends -- words that are orthographically similar but have completely different meanings in two languages -- as a possible approach to pinpoint the limitation of cross-lingual sense disambiguation in LLMs. We collect false friends in four language pairs, namely Indonesian-Malay, Indonesian-Tagalog, Chinese-Japanese, and English-German; and challenge LLMs to distinguish the use of them in context. In our analysis of various models, we observe they tend to be biased toward higher-resource languages. We also propose new metrics for quantifying the cross-lingual sense bias and comprehension based on our benchmark. Our work contributes to developing more diverse and inclusive language modeling, promoting fairer access for the wider multilingual community.

Vision Language Models (VLMs) often struggle with culture-specific knowledge, particularly in languages other than English and in underrepresented cultural contexts. To evaluate their understanding of such knowledge, we introduce WorldCuisines, a massive-scale benchmark for multilingual and multicultural, visually grounded language understanding. This benchmark includes a visual question answering (VQA) dataset with text-image pairs across 30 languages and dialects, spanning 9 language families and featuring over 1 million data points, making it the largest multicultural VQA benchmark to date. It includes tasks for identifying dish names and their origins. We provide evaluation datasets in two sizes (12k and 60k instances) alongside a training dataset (1 million instances). Our findings show that while VLMs perform better with correct location context, they struggle with adversarial contexts and predicting specific regional cuisines and languages. To support future research, we release a knowledge base with annotated food entries and images along with the VQA data., Comment: Accepted by NAACL 2025

What factors contribute to the success of national teams in the Men's Ice Hockey World Championships and the Men's Ice Hockey World Junior Championships? This study examines whether hosting the tournament provides a home advantage; the influence of past tournament performances; the impact of players' physical characteristics such as height, weight, and age; and the value of experience from the World Championships compared to the NHL and other leagues. We employ a dynamic ranking model based on the Plackett-Luce distribution with time-varying strength parameters driven by the score. Additionally, we conduct a forecasting analysis to predict the probabilities of winning the tournament, earning a medal, and advancing to the playoff phase.

Non-negative matrix factorization (NMF) is a key technique for feature extraction and widely used in source separation. However, existing algorithms may converge to poor local minima, or to one of several minima with similar objective value but differing feature parametrizations. Here we show that some of these weaknesses may be mitigated by performing NMF in a higher-dimensional feature space and then iteratively combining components with an analytically-solvable pairwise merge strategy. Experimental results demonstrate our method helps non-ideal NMF solutions escape to better local optima and achieve greater consistency of the solutions. Despite these extra steps, our approach exhibits similar computational performance to established methods by reducing the occurrence of "plateau phenomenon" near saddle points. Moreover, the results also illustrate that our method is compatible with different NMF algorithms. Thus, this can be recommended as a preferred approach for most applications of NMF.

Non-negative matrix factorization (NMF) is an important tool in signal processing and widely used to separate mixed sources into their components. Algorithms for NMF require that the user choose the number of components in advance, and if the results are unsatisfying one typically needs to start again with a different number of components. To make NMF more interactive and incremental, here we introduce GSVD-NMF, a method that proposes new components based on the generalized singular value decomposition (GSVD) to address discrepancies between the initial under-complete NMF results and the SVD of the original matrix. Simulation and experimental results demonstrate that GSVD-NMF often effectively recovers multiple missing components in under-complete NMF, with the recovered NMF solutions frequently reaching better local optima. The results further show that GSVD-NMF is compatible with various NMF algorithms and that directly augmenting components is more efficient than rerunning NMF from scratch with additional components. By deliberately starting from under-complete NMF, GSVD-NMF has the potential to be a recommended approach for a range of general NMF applications.

Southeast Asia (SEA) is a region rich in linguistic diversity and cultural variety, with over 1,300 indigenous languages and a population of 671 million people. However, prevailing AI models suffer from a significant lack of representation of texts, images, and audio datasets from SEA, compromising the quality of AI models for SEA languages. Evaluating models for SEA languages is challenging due to the scarcity of high-quality datasets, compounded by the dominance of English training data, raising concerns about potential cultural misrepresentation. To address these challenges, we introduce SEACrowd, a collaborative initiative that consolidates a comprehensive resource hub that fills the resource gap by providing standardized corpora in nearly 1,000 SEA languages across three modalities. Through our SEACrowd benchmarks, we assess the quality of AI models on 36 indigenous languages across 13 tasks, offering valuable insights into the current AI landscape in SEA. Furthermore, we propose strategies to facilitate greater AI advancements, maximizing potential utility and resource equity for the future of AI in SEA., Comment: https://seacrowd.github.io/ Accepted in EMNLP 2024

Visual Question Answering (VQA) is an important task in multimodal AI, and it is often used to test the ability of vision-language models to understand and reason on knowledge present in both visual and textual data. However, most of the current VQA models use datasets that are primarily focused on English and a few major world languages, with images that are typically Western-centric. While recent efforts have tried to increase the number of languages covered on VQA datasets, they still lack diversity in low-resource languages. More importantly, although these datasets often extend their linguistic range via translation or some other approaches, they usually keep images the same, resulting in narrow cultural representation. To address these limitations, we construct CVQA, a new Culturally-diverse multilingual Visual Question Answering benchmark, designed to cover a rich set of languages and cultures, where we engage native speakers and cultural experts in the data collection process. As a result, CVQA includes culturally-driven images and questions from across 30 countries on four continents, covering 31 languages with 13 scripts, providing a total of 10k questions. We then benchmark several Multimodal Large Language Models (MLLMs) on CVQA, and show that the dataset is challenging for the current state-of-the-art models. This benchmark can serve as a probing evaluation suite for assessing the cultural capability and bias of multimodal models and hopefully encourage more research efforts toward increasing cultural awareness and linguistic diversity in this field., Comment: 38th Conference on Neural Information Processing Systems (NeurIPS 2024) Track on Datasets and Benchmarks

This study presented the first light curve analysis of the OP Boo and V0511 Cam binary stars, which was conducted in the frame of the Binary Systems of South and North (BSN) Project. Photometric ground-based observations were conducted with standard filters at two observatories in the Czech Republic. We computed a new ephemeris for each of the systems using our extracted times of minima, TESS data, and additional literature. Linear fits for O-C diagrams of both systems were considered using the Markov Chain Monte Carlo (MCMC) method. The light curves were analyzed using the Wilson-Devinney (WD) binary code combined with the Monte Carlo (MC) simulation. The light curve solutions of both target systems required a cold starspot. The absolute parameters of the systems were calculated by using a P-M parameter relationship. The positions of the systems were also depicted on the Hertzsprung-Russell (HR), P-L, logMtot-logJ0, and T-M diagrams. The second component in both systems is determined to be a more massive and hotter star. Therefore, it can be concluded that both systems are W-type contact binary systems., Comment: 5 Tables, 4 Figures, Astrophysics Journal

Research and development (R&D) of countries play a major role in a long-term development of the economy. We measure the R&D efficiency of all 28 member countries of the European Union in the years 2008--2014. Super-efficient data envelopment analysis (DEA) based on robustness of classification into efficient and inefficient units is adopted. We use the number of citations as output of basic research, the number of patents as output of applied research and R&D expenditures with manpower as inputs. To meet DEA assumptions and to capture R&D characteristics, we analyze a homogeneous sample of countries, adjust prices using purchasing power parity and consider time lag between inputs and outputs. We find that the efficiency of general R&D is higher for countries with higher GDP per capita. This relation also holds for specialized efficiencies of basic and applied research. However, it is much stronger for applied research suggesting its outputs are more easily distinguished and captured. Our findings are important in the evaluation of research and policy making.

The categorization of retail products is essential for the business decision-making process. It is a common practice to classify products based on their quantitative and qualitative characteristics. In this paper we use a purely data-driven approach. Our clustering of products is based exclusively on the customer behaviour. We propose a method for clustering retail products using market basket data. Our model is formulated as an optimization problem which is solved by a genetic algorithm. It is demonstrated on simulated data how our method behaves in different settings. The application using real data from a Czech drugstore company shows that our method leads to similar results in comparison with the classification by experts. The number of clusters is a parameter of our algorithm. We demonstrate that if more clusters are allowed than the original number of categories is, the method yields additional information about the structure of the product categorization.

Generalized autoregressive score (GAS) models are a class of observation-driven time series models that employ the score to dynamically update time-varying parameters of the underlying probability distribution. GAS models have been extensively studied and numerous variants have been proposed in the literature to accommodate diverse data types and probability distributions. This paper introduces the gasmodel package, which has been designed to facilitate the estimation, forecasting, and simulation of a wide range of GAS models. The package provides a rich selection of distributions, offers flexible options for specifying dynamics, and allows to incorporate exogenous variables. Model estimation utilizes the maximum likelihood method.

Large language models (LLMs) show remarkable human-like capability in various domains and languages. However, a notable quality gap arises in low-resource languages, e.g., Indonesian indigenous languages, rendering them ineffective and inefficient in such linguistic contexts. To bridge this quality gap, we introduce Cendol, a collection of Indonesian LLMs encompassing both decoder-only and encoder-decoder architectures across a range of model sizes. We highlight Cendol's effectiveness across a diverse array of tasks, attaining 20% improvement, and demonstrate its capability to generalize to unseen tasks and indigenous languages of Indonesia. Furthermore, Cendol models showcase improved human favorability despite their limitations in capturing indigenous knowledge and cultural values in Indonesia. In addition, we discuss the shortcomings of parameter-efficient tunings, such as LoRA, for language adaptation. Alternatively, we propose the usage of vocabulary adaptation to enhance efficiency. Lastly, we evaluate the safety of Cendol and showcase that safety in pre-training in one language such as English is transferable to low-resource languages, such as Indonesian, even without RLHF and safety fine-tuning., Comment: Cendol models are released under Apache 2.0 license and will be made publicly available soon

In-context learning (ICL) empowers large language models (LLMs) to perform diverse tasks in underrepresented languages using only short in-context information, offering a crucial avenue for narrowing the gap between high-resource and low-resource languages. Nonetheless, there is only a handful of works explored ICL for low-resource languages with most of them focusing on relatively high-resource languages, such as French and Spanish. In this work, we extensively study ICL and its cross-lingual variation (X-ICL) on 25 low-resource and 7 relatively higher-resource languages. Our study not only assesses the effectiveness of ICL with LLMs in low-resource languages but also identifies the shortcomings of in-context label alignment, and introduces a more effective alternative: query alignment. Moreover, we provide valuable insights into various facets of ICL for low-resource languages. Our study concludes the significance of few-shot in-context information on enhancing the low-resource understanding quality of LLMs through semantically relevant information by closing the language gap in the target language and aligning the semantics between the targeted low-resource and the high-resource language that the model is proficient in. Our work highlights the importance of advancing ICL research, particularly for low-resource languages. Our code is publicly released at https://github.com/SamuelCahyawijaya/in-context-alignment

We introduce and study a new type of compactness principle for strong logics that, roughly speaking, infers the consistency of a theory from the consistency of its small fragments in certain outer models of the set-theoretic universe. We refer to this type of compactness property as outward compactness, and we show that instances of this type of principle for second-order logic can be used to characterize various large cardinal notions between measurability and extendibility, directly generalizing a classical result of Magidor that characterizes extendible cardinals as the strong compactness cardinals of second-order logic. In addition, we generalize a result of Makowsky that shows that Vop\v{e}nka's Principle is equivalent to the existence of compactness cardinals for all abstract logics by characterizing the principle "Ord is Woodin" through outward compactness properties of abstract logics.

This study intends to build smart water irrigation for rice farming using IoT and microcontroller devices with solar panel support. The system demonstrates the capabilities of automated irrigation by reducing physical labor through smart monitoring of the temperature, soil moisture, and humidity using multiple sensors. This study uses an agile methodology as it is suitable for reiterative operation for the development of the prototype. The mean result for the interpretation of data gathered for the systems' adaptability and flexibility is 4.32. The researchers were able to develop smart water irrigation for rice farming using IoT and microcontroller devices with solar panel support and the respondents also agreed that Smart water irrigation for rice farming using IoT and microcontroller devices with solar panel support is practical and valuable. A decision support system is recommended that can analyze data collected from IoT sensors and provide further recommendations. Based on the results, it is also suggested that future researchers use drip irrigation, instead of flood irrigation. Smart water irrigation has the potential to revolutionize agriculture, enhance environmental sustainability, and address pressing global challenges related to water resources and food security. These implications highlight the importance of continued research and innovation in this field., Comment: 14 pages, 5 figures, 1 table

Light-induced ion demixing in mixed-halide perovskites is simulated numerically using the phenomenological Cahn-Hilliard equation. In the model we consider the energy of local elastic deformation as well as the contribution of free carriers, assuming both the polaron and the bandgap-fluctuation models. The simulation shows that elastic deformation suppresses the demixing while free carriers promote it, however both effects lead to different ion distributions. The free-carrier-induced demixing appears only for larger starting random fluctuations of the ion concentration., Comment: 11 pages, 10 figures

Inference, especially those derived from inductive processes, is a crucial component in our conversation to complement the information implicitly or explicitly conveyed by a speaker. While recent large language models show remarkable advances in inference tasks, their performance in inductive reasoning, where not all information is present in the context, is far behind deductive reasoning. In this paper, we analyze the behavior of the models based on the task difficulty defined by the semantic information gap -- which distinguishes inductive and deductive reasoning (Johnson-Laird, 1988, 1993). Our analysis reveals that the disparity in information between dialogue contexts and desired inferences poses a significant challenge to the inductive inference process. To mitigate this information gap, we investigate a contrastive learning approach by feeding negative samples. Our experiments suggest negative samples help models understand what is wrong and improve their inference generations., Comment: Accepted to EMNLP2023

Large language models (LLMs) have been used for diverse tasks in natural language processing (NLP), yet remain under-explored for task-oriented dialogue systems (TODS), especially for end-to-end TODS. We present InstructTODS, a novel off-the-shelf framework for zero-shot end-to-end task-oriented dialogue systems that can adapt to diverse domains without fine-tuning. By leveraging LLMs, InstructTODS generates a proxy belief state that seamlessly translates user intentions into dynamic queries for efficient interaction with any KB. Our extensive experiments demonstrate that InstructTODS achieves comparable performance to fully fine-tuned TODS in guiding dialogues to successful completion without prior knowledge or task-specific data. Furthermore, a rigorous human evaluation of end-to-end TODS shows that InstructTODS produces dialogue responses that notably outperform both the gold responses and the state-of-the-art TODS in terms of helpfulness, informativeness, and humanness. Moreover, the effectiveness of LLMs in TODS is further supported by our comprehensive evaluations on TODS subtasks: dialogue state tracking, intent classification, and response generation. Code and implementations could be found here https://github.com/WillyHC22/InstructTODS/

Object hallucination poses a significant challenge in vision-language (VL) models, often leading to the generation of nonsensical or unfaithful responses with non-existent objects. However, the absence of a general measurement for evaluating object hallucination in VL models has hindered our understanding and ability to mitigate this issue. In this work, we present NOPE (Negative Object Presence Evaluation), a novel benchmark designed to assess object hallucination in VL models through visual question answering (VQA). We propose a cost-effective and scalable approach utilizing large language models to generate 29.5k synthetic negative pronoun (NegP) data of high quality for NOPE. We extensively investigate the performance of 10 state-of-the-art VL models in discerning the non-existence of objects in visual questions, where the ground truth answers are denoted as NegP (e.g., "none"). Additionally, we evaluate their standard performance on visual questions on 9 other VQA datasets. Through our experiments, we demonstrate that no VL model is immune to the vulnerability of object hallucination, as all models achieve accuracy below 10\% on NegP. Furthermore, we uncover that lexically diverse visual questions, question types with large scopes, and scene-relevant objects capitalize the risk of object hallucination in VL models., Comment: Published in ALVR Workshop at ACL 2024

In recent years, the rapid advancement of machine learning (ML) models, particularly transformer-based pre-trained models, has revolutionized Natural Language Processing (NLP) and Computer Vision (CV) fields. However, researchers have discovered that these models can inadvertently capture and reinforce social biases present in their training datasets, leading to potential social harms, such as uneven resource allocation and unfair representation of specific social groups. Addressing these biases and ensuring fairness in artificial intelligence (AI) systems has become a critical concern in the ML community. The recent introduction of pre-trained vision-and-language (VL) models in the emerging multimodal field demands attention to the potential social biases present in these models as well. Although VL models are susceptible to social bias, there is a limited understanding compared to the extensive discussions on bias in NLP and CV. This survey aims to provide researchers with a high-level insight into the similarities and differences of social bias studies in pre-trained models across NLP, CV, and VL. By examining these perspectives, the survey aims to offer valuable guidelines on how to approach and mitigate social bias in both unimodal and multimodal settings. The findings and recommendations presented here can benefit the ML community, fostering the development of fairer and non-biased AI models in various applications and research endeavors.

Grounding dialogue response generation on external knowledge is proposed to produce informative and engaging responses. However, current knowledge-grounded dialogue (KGD) systems often fail to align the generated responses with human-preferred qualities due to several issues like hallucination and the lack of coherence. Upon analyzing multiple language model generations, we observe the presence of alternative generated responses within a single decoding process. These alternative responses are more faithful and exhibit a comparable or higher level of relevance to prior conversational turns compared to the optimal responses prioritized by the decoding processes. To address these challenges and driven by these observations, we propose Polished \& Informed Candidate Scoring (PICK), a generation re-scoring framework that empowers models to generate faithful and relevant responses without requiring additional labeled data or model tuning. Through comprehensive automatic and human evaluations, we demonstrate the effectiveness of PICK in generating responses that are more faithful while keeping them relevant to the dialogue history. Furthermore, PICK consistently improves the system's performance with both oracle and retrieved knowledge in all decoding strategies. We provide the detailed implementation in https://github.com/bryanwilie/pick .

Democratizing access to natural language processing (NLP) technology is crucial, especially for underrepresented and extremely low-resource languages. Previous research has focused on developing labeled and unlabeled corpora for these languages through online scraping and document translation. While these methods have proven effective and cost-efficient, we have identified limitations in the resulting corpora, including a lack of lexical diversity and cultural relevance to local communities. To address this gap, we conduct a case study on Indonesian local languages. We compare the effectiveness of online scraping, human translation, and paragraph writing by native speakers in constructing datasets. Our findings demonstrate that datasets generated through paragraph writing by native speakers exhibit superior quality in terms of lexical diversity and cultural content. In addition, we present the \datasetname{} benchmark, encompassing 12 underrepresented and extremely low-resource languages spoken by millions of individuals in Indonesia. Our empirical experiment results using existing multilingual large language models conclude the need to extend these models to more underrepresented languages. We release the NusaWrites dataset at https://github.com/IndoNLP/nusa-writes.

CrI$_3$ represents one of the most important van der Waals systems on the route to understanding two-dimensional magnetic phenomena. Being arranged in a specific layered structure it also provides a unique opportunity to investigate structural transformations in dimension-confined systems. CrI$_3$ is dimorphic and possesses a higher symmetry low-temperature phase, which is quite uncommon. It contrasts with vanadium trihalides which show a higher symmetry high-temperature. An explanation of this distinct behavior together with a large cycle-dependent transition hysteresis is still an open question. Our low-temperature X-ray diffraction study conducted on CrI$_3$ single crystals complemented by magnetization and specific heat measurements was focused mainly on specific features of the structural transition during cooling. Our results manifest that the structural transition during cooling relates to the formation of structural domains despite the lower symmetry structure transforming to a higher symmetry one. We propose that these domains could control the transition temperature and also the size of thermal hysteresis.

An alternative approach for the panel second stage of data envelopment analysis (DEA) is presented in this paper. Instead of efficiency scores, we propose to model rankings in the second stage using a dynamic ranking model in the score-driven framework. We argue that this approach is suitable to complement traditional panel regression as a robustness check. To demonstrate the proposed approach, we determine research efficiency of higher education systems at country level by examining scientific publications and analyze its relation to good governance. The proposed approach confirms positive relation to the Voice and Accountability indicator, as found by the standard panel linear regression, while suggesting caution regarding the Government Effectiveness indicator.

Speech emotion recognition plays a crucial role in human-computer interactions. However, most speech emotion recognition research is biased toward English-speaking adults, which hinders its applicability to other demographic groups in different languages and age groups. In this work, we analyze the transferability of emotion recognition across three different languages--English, Mandarin Chinese, and Cantonese; and 2 different age groups--adults and the elderly. To conduct the experiment, we develop an English-Mandarin speech emotion benchmark for adults and the elderly, BiMotion, and a Cantonese speech emotion dataset, YueMotion. This study concludes that different language and age groups require specific speech features, thus making cross-lingual inference an unsuitable method. However, cross-group data augmentation is still beneficial to regularize the model, with linguistic distance being a significant influence on cross-lingual transferability. We release publicly release our code at https://github.com/HLTCHKUST/elderly_ser., Comment: Accepted in INTERSPEECH 2023

Large language models (LLMs) that are tuned with instructions have demonstrated remarkable capabilities in various tasks and languages. However, their ability to generalize to underrepresented languages is limited due to the scarcity of available data. Additionally, directly adapting new languages to instruction-tuned LLMs can result in catastrophic forgetting, which leads to the loss of multitasking ability. To address this issue, we propose InstructAlign which uses continual crosslingual instruction tuning to enable LLMs to align new unseen languages with previously learned high-resource languages. Our results demonstrate the effectiveness of InstructAlign in enabling the model to understand low-resource languages with limited parallel data while preventing catastrophic forgetting. Our work contributes to the advancement of language adaptation methods, particularly for adapting instruction-tuned LLMs to underrepresented languages. Our code is released on https://github.com/HLTCHKUST/InstructAlign

We study several properties of equi-Baire 1 families of functions between metric spaces. We consider the related equi-Lebesgue property for such families. We examine the behaviour of equi-Baire 1 and equi-Lebesgue families with respect to pointwise and uniform convergence. In particular, we obtain a criterion for a choice of a uniformly convergent subsequence from a sequence of functions that form an equi-Baire 1 family, which solves a problem posed in [3]. Finally, we discuss the notion of equi-cliquishness and relations between equi-Baire 1 families and sets of equi-continuity points.

While code-mixing is a common linguistic practice in many parts of the world, collecting high-quality and low-cost code-mixed data remains a challenge for natural language processing (NLP) research. The recent proliferation of Large Language Models (LLMs) compels one to ask: how capable are these systems in generating code-mixed data? In this paper, we explore prompting multilingual LLMs in a zero-shot manner to generate code-mixed data for seven languages in South East Asia (SEA), namely Indonesian, Malay, Chinese, Tagalog, Vietnamese, Tamil, and Singlish. We find that publicly available multilingual instruction-tuned models such as BLOOMZ and Flan-T5-XXL are incapable of producing texts with phrases or clauses from different languages. ChatGPT exhibits inconsistent capabilities in generating code-mixed texts, wherein its performance varies depending on the prompt template and language pairing. For instance, ChatGPT generates fluent and natural Singlish texts (an English-based creole spoken in Singapore), but for English-Tamil language pair, the system mostly produces grammatically incorrect or semantically meaningless utterances. Furthermore, it may erroneously introduce languages not specified in the prompt. Based on our investigation, existing multilingual LLMs exhibit a wide range of proficiency in code-mixed data generation for SEA languages. As such, we advise against using LLMs in this context without extensive human checks., Comment: Updating Authors

The demand for multimodal dialogue systems has been rising in various domains, emphasizing the importance of interpreting multimodal inputs from conversational and situational contexts. We explore three methods to tackle this problem and evaluate them on the largest situated dialogue dataset, SIMMC 2.1. Our best method, scene-dialogue alignment, improves the performance by ~20% F1-score compared to the SIMMC 2.1 baselines. We provide analysis and discussion regarding the limitation of our methods and the potential directions for future works. Our code is publicly available at https://github.com/holylovenia/multimodal-object-identification., Comment: Accepted at EACL SRW 2023

This paper proposes a framework for quantitatively evaluating interactive LLMs such as ChatGPT using publicly available data sets. We carry out an extensive technical evaluation of ChatGPT using 23 data sets covering 8 different common NLP application tasks. We evaluate the multitask, multilingual and multi-modal aspects of ChatGPT based on these data sets and a newly designed multimodal dataset. We find that ChatGPT outperforms LLMs with zero-shot learning on most tasks and even outperforms fine-tuned models on some tasks. We find that it is better at understanding non-Latin script languages than generating them. It is able to generate multimodal content from textual prompts, via an intermediate code generation step. Moreover, we find that ChatGPT is 63.41% accurate on average in 10 different reasoning categories under logical reasoning, non-textual reasoning, and commonsense reasoning, hence making it an unreliable reasoner. It is, for example, better at deductive than inductive reasoning. ChatGPT suffers from hallucination problems like other LLMs and it generates more extrinsic hallucinations from its parametric memory as it does not have access to an external knowledge base. Finally, the interactive feature of ChatGPT enables human collaboration with the underlying LLM to improve its performance, i.e, 8% ROUGE-1 on summarization and 2% ChrF++ on machine translation, in a multi-turn "prompt engineering" fashion. We also release codebase for evaluation set extraction., Comment: 45 pages, AACL 2022

We present NusaCrowd, a collaborative initiative to collect and unify existing resources for Indonesian languages, including opening access to previously non-public resources. Through this initiative, we have brought together 137 datasets and 118 standardized data loaders. The quality of the datasets has been assessed manually and automatically, and their value is demonstrated through multiple experiments. NusaCrowd's data collection enables the creation of the first zero-shot benchmarks for natural language understanding and generation in Indonesian and the local languages of Indonesia. Furthermore, NusaCrowd brings the creation of the first multilingual automatic speech recognition benchmark in Indonesian and the local languages of Indonesia. Our work strives to advance natural language processing (NLP) research for languages that are under-represented despite being widely spoken.

We develop a novel observation-driven model for high-frequency prices. We account for irregularly spaced observations, simultaneous transactions, discreteness of prices, and market microstructure noise. The relation between trade durations and price volatility, as well as intraday patterns of trade durations and price volatility, is captured using smoothing splines. The dynamic model is based on the zero-inflated Skellam distribution with time-varying volatility in a score-driven framework. Market microstructure noise is filtered by including a moving average component. The model is estimated by the maximum likelihood method. In an empirical study of the IBM stock, we demonstrate that the model provides a good fit to the data. Besides modeling intraday volatility, it can also be used to measure daily realized volatility.

Large pre-trained language models (LMs) have been widely adopted in biomedical and clinical domains, introducing many powerful LMs such as bio-lm and BioELECTRA. However, the applicability of these methods to real clinical use cases is hindered, due to the limitation of pre-trained LMs in processing long textual data with thousands of words, which is a common length for a clinical note. In this work, we explore long-range adaptation from such LMs with Longformer, allowing the LMs to capture longer clinical notes context. We conduct experiments on three n2c2 challenges datasets and a longitudinal clinical dataset from Hong Kong Hospital Authority electronic health record (EHR) system to show the effectiveness and generalizability of this concept, achieving 10\% F1-score improvement. Based on our experiments, we conclude that capturing a longer clinical note interval is beneficial to the model performance, but there are different cut-off intervals to achieve the optimal performance for different target variables. Our code is available at https://github.com/HLTCHKUST/long-biomedical-model.

Audio-based pornographic detection enables efficient adult content filtering without sacrificing performance by exploiting distinct spectral characteristics. To improve it, we explore pornographic sound modeling based on different neural architectures and acoustic features. We find that CNN trained on log mel spectrogram achieves the best performance on Pornography-800 dataset. Our experiment results also show that log mel spectrogram allows better representations for the models to recognize pornographic sounds. Finally, to classify whole audio waveforms rather than segments, we employ voting segment-to-audio technique that yields the best audio-level detection results., Comment: Published in AudioMostly 2022, ACM

Generating a short story out of an image is arduous. Unlike image captioning, story generation from an image poses multiple challenges: preserving the story coherence, appropriately assessing the quality of the story, steering the generated story into a certain style, and addressing the scarcity of image-story pair reference datasets limiting supervision during training. In this work, we introduce Plug-and-Play Story Teller (PPST) and improve image-to-story generation by: 1) alleviating the data scarcity problem by incorporating large pre-trained models, namely CLIP and GPT-2, to facilitate a fluent image-to-text generation with minimal supervision, and 2) enabling a more style-relevant generation by incorporating stylistic adapters to control the story generation. We conduct image-to-story generation experiments with non-styled, romance-styled, and action-styled PPST approaches and compare our generated stories with those of previous work over three aspects, i.e., story coherence, image-story relevance, and style fitness, using both automatic and human evaluation. The results show that PPST improves story coherence and has better image-story relevance, but has yet to be adequately stylistic., Comment: Accepted in LaTeCH-CLfL 2022 (6th Joint SIGHUM Workshop on Computational Linguistics for Cultural Heritage, Social Sciences, Humanities and Literature), COLING 2022

At the center of the underlying issues that halt Indonesian natural language processing (NLP) research advancement, we find data scarcity. Resources in Indonesian languages, especially the local ones, are extremely scarce and underrepresented. Many Indonesian researchers do not publish their dataset. Furthermore, the few public datasets that we have are scattered across different platforms, thus makes performing reproducible and data-centric research in Indonesian NLP even more arduous. Rising to this challenge, we initiate the first Indonesian NLP crowdsourcing effort, NusaCrowd. NusaCrowd strives to provide the largest datasheets aggregation with standardized data loading for NLP tasks in all Indonesian languages. By enabling open and centralized access to Indonesian NLP resources, we hope NusaCrowd can tackle the data scarcity problem hindering NLP progress in Indonesia and bring NLP practitioners to move towards collaboration.

We investigate a variety of cut and choose games, their relationship with (generic) large cardinals, and show that they can be used to characterize a number of properties of ideals and of partial orders: certain notions of distributivity, strategic closure, and precipitousness., Comment: 31 pages