Hugin Expert (Denmark)

Data from psychophysiological measures can offer new insight into control room operators’ behaviour, cognition, and mental workload status. This is particularly helpful when assessing capacity to respond to critical plant conditions such as alarm response scenarios. However, wearable tools such as eye tracking and electroencephalography caps can be perceived as intrusive and unsuitable for daily operations. Therefore, this article examines the potential of using real-time data from process and operator-system interactions during abnormal scenarios, recorded and retrieved from the distributed control system’s historian or process log, to provide insight into operator behaviour and predict their response outcomes without intruding on daily tasks. Data for this study were obtained from a design of experiment using a formaldehyde production plant simulator and four human-in-the-loop support configurations. A comparison between configurations in terms of both behaviour and performance is presented. Then, a step-wise logistic regression and a Bayesian network model were used to predict operator error. The results identified predictive metrics, discussed in terms of their value as precursors of overall system performance in alarm response scenarios. Knowledge of relevant and predictive behavioural metrics accessible in real time can better equip decision-makers to predict outcomes and provide timely support measures for operators.

Cut-in maneuvers in high-speed traffic pose critical challenges that can lead to abrupt braking and collisions, necessitating safe and efficient lane change strategies. We propose a Dynamic Bayesian Network (DBN) framework to integrate lateral evidence with safety assessment models, thereby predicting lane changes and ensuring safe cut-in maneuvers effectively. Our proposed framework comprises three key probabilistic hypotheses (lateral evidence, lateral safety, and longitudinal safety) that facilitate the decision-making process through dynamic data processing and assessments of vehicle positions, lateral velocities, relative distance, and Time-to-Collision (TTC) computations. The DBN model's performance compared with other conventional approaches demonstrates superior performance in crash reduction, especially in critical high-speed scenarios, while maintaining a competitive performance in low-speed scenarios. This paves the way for robust, scalable, and efficient safety validation in automated driving systems.

v4.9.2: Patch release Tpu tie weights #13030 (@sgugger) ONNX fixes & examples: #13048, #13049, #13028, #13014, #12911, (@mfuntowicz, @michaelbenayoun, @LysandreJik) Fix push_to_hub for TPUs #12895 (@sgugger)

New model architecture: DistilBERT Adding Huggingface's new transformer architecture, <strong>DistilBERT</strong> described in Smaller, faster, cheaper, lighter: Introducing DistilBERT, a distilled version of BERT by Victor Sanh, Lysandre Debut and Thomas Wolf. This new model architecture comes with two pretrained checkpoints: <code>distilbert-base-uncased</code>: the base DistilBert model <code>distilbert-base-uncased-distilled-squad</code>: DistilBert model fine-tuned with distillation on SQuAD. An awaited new pretrained checkpoint: GPT-2 large (774M parameters) The third OpenAI GPT-2 checkpoint (GPT-2 large) is available in the library under the shortcut name <code>gpt2-large</code>: 774M parameters, 36 layers, and 20 heads. New XLM multilingual pretrained checkpoints in 17 and 100 languages We have added two new XLM models in 17 and 100 languages which obtain better performance than multilingual BERT on the XNLI cross-lingual classification task. New dependency: <code>sacremoses</code> Support for XLM is improved by carefully reproducing the original tokenization workflow (work by @shijie-wu in #1092). We now rely on <code>sacremoses</code>, a python port of Moses tokenizer, truecaser and normalizer by @alvations, for XLM word tokenization. In a few languages (Thai, Japanese and Chinese) XLM tokenizer will require additional dependencies. These additional dependencies are optional at the library level. Using XLM tokenizer in these languages without the additional dependency will raise an error message with installation instructions. The additional optional dependencies are: pythainlp: Thai tokenizer kytea: Japanese tokenizer, wrapper of KyTea (Need external C++ compilation), used by the newly release XLM-17 &amp; XLM-100 jieba: Chinese tokenizer * * XLM used Stanford Segmenter. However, the wrapper (nltk.tokenize.stanford_segmenter) are slow due to JVM overhead, and it will be deprecated. Jieba is a lot faster and pip-installable. But there is some mismatch with the Stanford Segmenter. A workaround could be having an argument to allow users to segment the sentence by themselves and bypass the segmenter. As a reference, I also include nltk.tokenize.stanford_segmenter in this PR. Bug fixes and improvements to the library modules Bertology script has seen major improvements (@tuvuumass ) Iterative tokenization now faster and accept arbitrary numbers of added tokens (@samvelyan) Added RoBERTa to AutoModels and AutoTokenizers (@LysandreJik ) Added GPT-2 Large 774M model (@thomwolf ) Added language model fine-tuning with GPT/GPT-2 (CLM), BERT/RoBERTa (MLM) (@LysandreJik @thomwolf ) Multi-GPU training has been patched (@FeiWang96 ) Scripts are updated to reflect Pytorch 1.1.0 changes (scheduler, optimizer) (@Morizeyao, @adai183 ) Updated the in-depth BERT fine-tuning scripts to <code>pytorch-transformers</code> (@Morizeyao ) Models saved with pruned heads are now saved and reloaded correctly (implemented for GPT, GPT-2, BERT, RoBERTa, XLM) (@LysandreJik @thomwolf) Add <code>proxies</code> and <code>force_download</code> options to <code>from_pretrained()</code> method to be able to use proxies and update cached models/tokenizers (@thomwolf) Add shortcut to each special tokens with <code>_id</code> properties (e.g. <code>tokenizer.cls_token_id</code> for the id in the vocabulary of <code>tokenizer.cls_token</code>) (@thomwolf) Fix GPT2 and RoBERTa tokenizer so that sentences to be tokenized always begins with at least one space (see note by fairseq authors) (@thomwolf) Fix and clean up byte-level BPE tests (@thomwolf) Update the test classes for OpenAI GPT and GPT-2 so that these models are tested against common tests. (@LysandreJik ) Fix a warning raised when the decode method is called for a model with no <code>sep_token</code> like GPT-2 (@LysandreJik ) Updated the tokenizers saving method (@boy2000-007man) SpaCy tokenizers have been updated in the tokenizers (@GuillemGSubies ) Stable <code>EnvironmentErrors</code> have been added to utility files (@abhishekraok ) Fixed distributed barrier hang (@VictorSanh ) Encoding functions now return the input tokens instead of throwing an error when not implemented in child class (@LysandreJik ) Change layer norm code to PyTorch's native layer norm (@dhpollack) Improve tokenization of XLM for multilingual inputs (@shijie-wu) Add language input and access to language to id conversion in XLM tokenizer (@thomwolf) Add pretrained configuration properties for tokenizers with serialization logic (saving/reloading tokenizer configuration) (@thomwolf) Added new AutoModels: <code>AutoModelWithLMHead</code>, <code>AutoModelForSequenceClassification</code>, <code>AutoModelForQuestionAnswering</code> (@LysandreJik) Torch.hub is now based on AutoModels (@LysandreJik @thomwolf) Fix Transformer-XL attention mask dtype to be bool (@CrafterKolyan) Adding DistilBert model architecture and checpoints (@VictorSanh @LysandreJik @thomwolf) Fixes to DistilBert configuration and training script (@stefan-it) Fix XLNet attention mask for fp16 (@ziliwang) Documentation auto-deploy (@LysandreJik) Fix to add a tuple of tokens (@epwalsh) Update fp16 apex implmentation in scripts (@anhnt170489) Fix XLNet bias resizing when adding/removing tokens (@LysandreJik) Fix tokenizer reloading in example scripts (@rabeehk) Fix byte-level decoding error when using added tokens (@thomwolf @LysandreJik) Fix epsilon value in RoBERTa pretrained checkpoints (@julien-c)

In this paper, we exploit the symmetry of independence in the implementation of d-separation. We show that it can matter whether the search is conducted from start to goal or vice versa. Analysis reveals it is preferable to approach observed v-structure nodes from the bottom. Hence, a measure, called depth, is suggested to decide whether the search should run from start to goal or from goal to start. One salient feature is that depth can be computed during a pruning optimization step widely implemented. An empirical comparison is conducted against a clever implementation of d-separation. The experimental results are promising in two aspects. The effectiveness of our method increases with network size, as well as with the amount of observed evidence, culminating with an average time savings of 9% in the 9 largest BNs used in our experiments.

In this paper, we consider the application of object-oriented Bayesian networks to failure diagnostics in manufacturing systems and continuous model improvement based on operational data. The analysis is based on an object-oriented Bayesian network developed for failure diagnostics of a one-dimensional pick-and-place industrial robot developed by IEF-Werner GmbH. We consider four learning algorithms (batch Expectation-Maximization (EM), incremental EM, Online EM and fractional updating) for parameter updating in the object-oriented Bayesian network using a real operational dataset. Also, we evaluate the performance of the considered algorithms on a dataset generated from the model to determine which algorithm is best suited for recovering the underlying generating distribution. The object-oriented Bayesian network has been integrated into both the control software of the robot as well as into a software architecture that supports diagnostic and prognostic capabilities of devices in manufacturing systems. We evaluate the time performance of the architecture to determine the feasibility of on-line learning from operational data using each of the four algorithms.

SUMMARYThe control ofCampylobacterin poultry is considered a public health priority and some intervention strategies have been implemented in Denmark. Nonetheless,Campylobacterinfection in poultry can still be considerable particularly during the summer when the most promisingCampylobactercontrol strategy seems to be the use of fly screens. The use of cost-effective vaccines againstCampylobacteris also desirable. In order to controlCampylobacter, poultry producers need to make crucial decisions under conditions of uncertainty. With the aim of assisting poultry producers in decision making regardingCampylobactercontrol strategies, the objective of the present study was to produce a decision support system that integrated knowledge and used a Bayesian approach to handle uncertainty. This decision support system integrated epidemiological data, microbiological considerations, financial information and potential control strategies (the use of fly screens and hypothetical vaccines). In conclusion, results from model and sensitivity analyses indicated that the financial variables (cost–benefit functions) and the effectiveness of the different control strategies drove the results.

This paper presents a business case for the development and operation of EV (Electric Vehicle) fast charging stations on the Danish motorway network. The business case is based upon an analysis of the EV models available on the Danish market, existing high power recharging solutions and a projection of market uptake of EVs in Denmark. The business case investigates two scenarios: A low scenario where 5 % of an EV’s charging demand is covered by fast charging stations and a high scenario where fast charging stations cover 10 % of an EV’s charging requirements. The two scenarios are examined for both a 5-year and a 10-year period. The business case shows that fast charging stations throughout a period of 10 years may result in a significant return of investments whereas investments for a period of only 5 years is too risky.