DOCSLIB.ORG

Anton Nikolaev and Nick Glaser

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Anton Nikolaev and Nick Glaser Tensorflow2.0 Question Answering Anton Nikolaev, Nick Glaser ICS 661 - Final Report 1. Introduction applicable, sections from Wikipedia articles Natural language processing (NLP) containing the answer. In contrast to some is one of the domains where the emergence other QA datasets, NQ also provides of deep learning (DL) has had the largest answer candidates for each question as well impact, improving performance across as a context level indicator. The candidates almost the entire spectrum of NLP. contain the indices representing the One of the kinds of problems that are respective start and end tokens for each currently being solved by DL researchers answer. The context indicator is a binary are reading comprehension/ question value that signals whether a given answering (QA) problems. For our project, candidate answer is also contained within we joined a kaggle competition based on a another candidate (nested) or whether it is novel QA dataset provided by Google the only candidate containing the specific Research titled Natural Questions (NQ). Our passage (top-level). This additional goal was to evaluate the performance of information can help improve model some of the current state-of-the-art NLP accuracy after the training stage, but is not architectures on this dataset. traditionally used during training itself. Overall, the entire dataset contains 2. Problem and Dataset about 300,000 training examples as well as The goal of the QA task is just under 8000 test examples that are essentially two-fold: the algorithm is ultimately used to evaluate our model’s provided with a text passage and a performance on kaggle. corresponding question from an arbitrary An example question from the training data domain with the goal of first understanding is given in Appendix A. For more details on the question and subsequently providing the the dataset, visit: appropriate answer (given that one exists) https://github.com/google-resea extracted from the text passage at hand. rch-datasets/natural-questions For the problem-formulation in the kaggle challenge, the answer could potentially 2.1 Metrics come in four formats: as a long answer, For the competition, correct answers usually being about a paragraph in length; must have the correct answer type selected as a short answer, being no longer than a and, where applicable, must also provide sentence and as short as a few words; as a the exactly correct answer span. If the yes/no answer, which is treated as a correct answer type is selected but the sub-type of the short answer; or, lastly, answer span is incorrect, that will still be there could be no answer in the text which considered a misclassification. the algorithm would also have to correctly Overall, the relevant metric is micro-F1 identify. between all answer types. F1 is the The dataset itself consists of google harmonic mean between the precision and queries taken from actual users and, if recall of the model. If the model has multiple answer types, we instead take the micro complicated. These features however, average of the precision and recall metrics comes with the consequence that there is for each answer type. So, if we were to no way to support infinitely long input have two answer types and wanted to get sequences and instead the micro-F1, we would compute the input-sequence-length has to be constant following: and requires padding/ truncating the data. Nevertheless, empirically it is evident that tp1+tp2 this trade-off is overall beneficial to micro average precision = tp +tp +fp +fp 1 2 1 2 performance. tp +tp micro average recall = 1 2 While we plan on trying other architectures tp1+tp2+fn1+fn2 in the future, thus far, our results are limited to BERT [2]. In its essence, BERT, like And, thus, finally, the micro-F1 would just many of the other models, is just a very be: deep network with many subsequent m.a. precision · m.a. recall Transformer layers and many attention micro − F 1 = 2 · m.a. precision + m.a. recall heads per layer. What makes BERT and other large language models so powerful aside from their incredibly large number of trainable parameters (110M for the small 3. Background BERT model), is the very scalable initial The Transformer architecture training procedure. Language models (Figure 1) has two main distinctive features, heavily depend on pretraining on large ​ its bidirectionality and its multi-headed corpora. attention mechanism, and in concept is similar to other encoder-decoder networks [1]. While it is beyond the scope of this report to fully explore the architecture, it is worth highlighting how these features contribute to the impact the architecture has had in the past two years. Both, bidirectionality and multi-head attention, allow the model to establish more complicated dependencies than other commonly used models like RNNs would be able to and are thus better suited to model language data. RNNs, specifically the LSTMs [3] which are the most commonly-used variant of recurrent networks, are strongly limited by the way in which they sequentially process the data which doesn’t always make sense for Figure 1: Transformer Architecture [1] language, where relationships between ​ ​ components are often much more In the case of BERT, this is achieved in two and short answers, we need to have ways, the first of which is reminiscent of multiple outputs for the model. how neural bag-of-word embedding models Overall, we ended up with three outputs. are trained. The input consists of a The first two outputs contain logits paragraph of text with 15% of the words corresponding to each token of the input masked and the model is then required to paragraph. These can be understood as a predict those words from their context. value indicating how likely a given token is The second task is known as to be the start or end token of the correct next-sentence-prediction (NSP). In NSP, the answer to the question. This will be model is given a context paragraph and explained in more detail in the then a potential next sentence. It then has post-processing section. The last output is a to perform binary classification to indicate five-unit vector that indicates what type of whether that sentence truly follows the answer we expect for the input question given context or was simply picked at (short, long, yes, no, no-answer). random from the corpus, with both cases occurring about half of the time. This 4.2 Data Processing training task is perhaps more closely related Preprocessing has to be done to the QA problem that we are interested in. exactly as is detailed in the BERT paper [2] Due to the semi-supervised nature of these since the input format needs to be tasks, it is easy to scale this to very large consistent between the pre-training and training corpora, which are in BERTs case fine-tuning/ prediction phases. The most the Book Corpus (800M words) and all of notable preprocessing component is the English Wikipedia (2,500M words). tokenization of the paragraph, which is done by matching the tokens to a predefined 4. Our Approach vocabulary based on a greedy Currently, most QA datasets have longest-match-first approach. After their performance leaderboards dominated tokenization, the sequence is padded or by Transformer-based models (for a truncated to the length of 512 tokens. collection of datasets and leaderboards During fine-tuning, processing the model visit: outputs is still relatively consistent with what paperswithcode.com/task/questio is described in [2], however, in order to train n-answering). the answer-type output layer an additional ​ We set out to use some of the most loss-term has to be added to account for the prevalent of these architectures and adapt additional outputs. them for the specific formulation of this QA Handling the model outputs requires task. some more work because the competition was set up differently from common QA 4.1 Model tasks as mentioned above. We downloaded one of the publicly First, the logits are analyzed. The model is available pretrained BERT models and then trained to increase the start-end logit scores adapted and fine-tuned it for the QA task. based on the short answer span. So we Because we have to predict both the take the 20 highest scoring start and end answer type and start and indices of long logits respectively, and then create all possible pairs. From the pool of pairs, we ensemble their prediction to hopefully create select the viable candidates, meaning the a more robust overall model. start index is before the end index, and then With that being said, currently our sum their logit scores. The highest overall micro-F1 score is at 0.57, which, at the scoring candidate is then selected as a writing of this report, ranks 36th out of short answer. approximately 850 contestants. We are Next, the long answer is selected out of the confident that the addition of more models; pool of predefined candidates based on two proper answer selection, including proper criteria: it must contain the highest scoring yes/no-answer selection; and ensembling short answer, and it must be a top-level will help further improve that ranking. answer which prevents us from getting multiple matches. 6. Summary Ultimately, which answer type is chosen For sophisticated NLP tasks like depends on both the combined question answering, one of the biggest start-end-logit score as well as the outputs takeaways from this project is the realization in the answer-type vector. The scores are that the pre- and post-processing are some combined and whether they pass an of the largest parts of the puzzle.
Load more

Recommended publications
  • London Cries & Public Edifices
    >m ^Victoria %S COLLECTION OF VICTORIAN BOOKS AT BRIGHAM YOUNG UNIVERSITY Victorian 914.21 L533L 1851 3 1197 22902 7856 A,AA A ,' s 7rs a' lEn! 31113 rf K* I 'r X ^i W\lf' ^ J.eU ^W^3 mmm y<i mm§ ft Hftij •: :ii v^ ANDON431IE GRMMT am &U<2<3Slg,SORS TT© KEWBgRy A.KfD HARRIS *S) A SORNER OF1 3-AjWTT PAUL'S 6HUR6H-TARD, LONDON UPB Tfffi TOWfiR QT LONDON. A POTS & KETTLES TO MERaBELLOWiS TO MEND. POTS AND KETTLES TO MEND !—COPPER OR BRASS TO MEND ! The Tinker is swinging his fire-pot to make it burn, having placed his soldering-iron in it, and is proceeding to some corner or post, there to repair the saucepan he carries.—We commence with the most in- teresting edifice in our capital, THE TOWER OF LONDON; the fortress, the palace, and prison, in which so many events, connected with the history of our country, have transpired. The building with four towers in the centre is said to have been erected by William the Conqueror, and is the oldest part of the fortress. The small bell- tower in the front of our picture is that of the church of St. Peter's, (the tower being a parish itself,) on the Tower Green, erected in the reign of Edward I. Our view is taken from Tower Hill, near which was the scaffold on which so many have fallen. To the left of the picture stood the grand storehouse of William III., destroyed by fire, Nov. 1841. The Regalia is deposited here, and exhibited to the public, as is also the Horse Armoury.
    [Show full text]
  • The Construction of Northumberland House and the Patronage of Its Original Builder, Lord Henry Howard, 1603–14
    The Antiquaries Journal, 90, 2010,pp1 of 60 r The Society of Antiquaries of London, 2010 doi:10.1017⁄s0003581510000016 THE CONSTRUCTION OF NORTHUMBERLAND HOUSE AND THE PATRONAGE OF ITS ORIGINAL BUILDER, LORD HENRY HOWARD, 1603–14 Manolo Guerci Manolo Guerci, Kent School of Architecture, University of Kent, Marlowe Building, Canterbury CT27NR, UK. E-mail: [email protected] This paper affords a complete analysis of the construction of the original Northampton (later Northumberland) House in the Strand (demolished in 1874), which has never been fully investigated. It begins with an examination of the little-known architectural patronage of its builder, Lord Henry Howard, 1st Earl of Northampton from 1603, one of the most interesting figures of the early Stuart era. With reference to the building of the contemporary Salisbury House by Sir Robert Cecil, 1st Earl of Salisbury, the only other Strand palace to be built in the early seventeenth century, textual and visual evidence are closely investigated. A rediscovered eleva- tional drawing of the original front of Northampton House is also discussed. By associating it with other sources, such as the first inventory of the house (transcribed in the Appendix), the inside and outside of Northampton House as Henry Howard left it in 1614 are re-configured for the first time. Northumberland House was the greatest representative of the old aristocratic mansions on the Strand – the almost uninterrupted series of waterfront palaces and large gardens that stretched from Westminster to the City of London, the political and economic centres of the country, respectively. Northumberland House was also the only one to have survived into the age of photography.
    [Show full text]
  • Rapier House, Turnmill Street, London, EC I I'. L.Reruti,E
    REDHEADS ADVERTISING LTD. SOMMERVILLE & MILNE 21 Eldon LTD. Square, Newcastle -on -Tyne, 1. 216 Bothwell Street, Glasgow, C.2. Scotland. REX PUBLICITY SERVICE LTD. Director S Manager: J. Bruce Omand. 131-134 New Bond Street, London, W.I. THE W. J. SOUTHCOMBE ADVERTISING Telephone: 'Mayfair 7571. AGENCY LTD. T.Y. Executive: R. C. Granger. 167 Strand, oLndon, W.C.2. Telephone: RIPLEY, PRESTON & CO. LTD. Temple Bar 4273. Ludgate House, 107-111 Fleet Street, Lon- SPOTTISWOODE ADVERTISING LTD. don, E.C.4. Telephone: Central 8672. 34 Brook Street, London, W.1. Telephone: T.V. Executive: Walter A. Clare. Hyde Park 1221 MAX RITSON & PARTNERS LTD. T.I Lxecutive: W. J. Barter. 33 Henrietta Street, London, W.C.2. Tele- HENRY SQUIRE & CO. LTD. phone: Temple Bar 3861. Canada House, Norfolk Street, London, W.C. T.V. Executive: Roy G. Clark. 2. Telephone: Temple Bar 9093. ROE TELEVISION LTD. I I'. L.reruti,e: S. Windridge. (F. John Roe Ltd.) 20 St. Ann's Square, STEELE'S ADVERTISING SERVICE LTD. Manchester & 73 Grosvenor Street, London, 34 Brook Street, London, W.1. W.1. Telephone: Grosvenor 8228. T.Y. Executives: Graham Roe, Derek J. Roe. STEPHENS ADVERTISING SERVICE LTD. Clarendon House, 11-12 Clifford Street, ROLES & PARKER LTD. New Bond Street, London, W.1. Telephone: Rapier House, Turnmill Street, London, E.C. Hyde Park 1. 1641. Telephone: Clerkenwell 0545. Executive: E. W. R. T.V. Executive: H. T. Parker. Galley. STRAND PUBLICITY LTD. G. S. ROYDS LTD. 10 Stanhope Row, London, W.1. Telephone: 160 Piccadilly, London, W1.. Telephone: Grosvenor 1352. Hyde Park 8238.
    [Show full text]
  • Construction “Closing Date and Time”)
    UNCLASSIFIED A2. TITLE A1. DEPARTMENTAL REPRESENTATIVE London Chancery Consolidation Project, United Kingdom A3. SOLICITATION NUMBER A4.PROJECT NUMBER A5.DATE Ms. Jane Bird ARD-LDN-CONST-13094/A B-LDN-135 October 20th, 2013 Project Director (Chancery Consolidation Project) A6. RFPR DOCUMENTS High Commission of Canada 1. Request for Pre-qualification Responses (“RFPR”) title page 2. Submission Requirements and Evaluations (Section “I” – “SR” provisions) Macdonald House, 1 Grosvenor Square 3. General Instructions (Section “II” – “GI” provisions) London, United Kingdom 4. Questionnaire (Section “III” – “Q” provisions) In the event of discrepancies, inconsistencies or ambiguities of the wording of these Telephone : +44 (0) 207 004 6052 documents, the document that appears first on the above list shall prevail. E-mail: [email protected] A7. RESPONSE DELIVERY In order for the Response to be valid, it must be received no later than 14:00 (2:00 pm) EST on November 15th 2013 (Ottawa, Ontario, Canada) (“Closing Date” or Construction “Closing Date and Time”). Best Value Evaluation The Response is only to be emailed to the following address; [email protected] Solicitation#: ARD-LDN-CONST-13094/A Respondents should ensure that their name, address, Closing Date, and solicitation PHASE 1 – Pre-qualification number is clearly marked on their email. Failure to comply may result in the Response of a 2-Phase Procurement being declared non-compliant and rejected from further consideration. Process A8. RESPONSE Please note SR2. Respondents must meet mandatory requirements and must obtain a for minimum rating of ‘100’. Up to six (6) Respondents with the highest score will be invited to advance to Phase 2 of the procurement process.
    [Show full text]
  • An Examination of the Artist's Depiction of the City and Its Gardens 1745-1756
    Durham E-Theses Public and private space in Canaletto's London: An examination of the artist's depiction of the city and its gardens 1745-1756 Hudson, Ferne Olivia How to cite: Hudson, Ferne Olivia (2000) Public and private space in Canaletto's London: An examination of the artist's depiction of the city and its gardens 1745-1756, Durham theses, Durham University. Available at Durham E-Theses Online: http://etheses.dur.ac.uk/4252/ Use policy The full-text may be used and/or reproduced, and given to third parties in any format or medium, without prior permission or charge, for personal research or study, educational, or not-for-prot purposes provided that: • a full bibliographic reference is made to the original source • a link is made to the metadata record in Durham E-Theses • the full-text is not changed in any way The full-text must not be sold in any format or medium without the formal permission of the copyright holders. Please consult the full Durham E-Theses policy for further details. Academic Support Oce, Durham University, University Oce, Old Elvet, Durham DH1 3HP e-mail: [email protected] Tel: +44 0191 334 6107 http://etheses.dur.ac.uk 2 Public and Private Space in Canaletto's London. An Examination of the Artist's Depiction of the City and its Gardens 1745-1756. The copyright of this thesis rests with the author. No quotation from it should be published in any form, including Electronic and the Internet, without the author's prior written consent.
    [Show full text]
  • The Heart of the Empire
    The heart of the Empire A self-guided walk along the Strand ww.discoverin w gbrita in.o the stories of our rg lands discovered th cape rough w s alks 2 Contents Introduction 4 Route map 5 Practical information 6 Commentary 8 Credits 30 © The Royal Geographical Society with the Institute of British Geographers, London, 2015 Discovering Britain is a project of the Royal Geographical Society (with IBG) The digital and print maps used for Discovering Britain are licensed to the RGS-IBG from Ordnance Survey Cover image: Detail of South Africa House © Mike jackson RGS-IBG Discovering Britain 3 The heart of the Empire Discover London’s Strand and its imperial connections At its height, Britain’s Empire covered one-quarter of the Earth’s land area and one-third of the world’s population. It was the largest Empire in history. If the Empire’s beating heart was London, then The Strand was one of its major arteries. This mile- long street beside the River Thames was home to some of the Empire’s administrative, legal and commercial functions. The days of Empire are long gone but its legacy remains in the landscape. A walk down this modern London street is a fascinating journey through Britain’s imperial history. This walk was created in 2012 by Mike Jackson and Gary Gray, both Fellows of the Royal Geographical Society (with IBG). It was originally part of a series that explored how our towns and cities have been shaped for many centuries by some of the 206 participating nations in the 2012 Olympic and Paralympic Games.
    [Show full text]
  • The Jubilee Walkway. Section 5 of 5
    Transport for London. The Jubilee Walkway. Section 5 of 5. The Jubilee Loop (anti-clockwise). Start and finish: The Jubilee Walkway panel (on the south side of Trafalgar Square). Nearest station: Charing Cross . Section distance: 2 miles (3 kilometres). Introduction. This is a circular walk in the heart of London linking the institutions of The Monarchy and the Government. This section of the Jubilee Walkway connects with section one at Trafalgar Square and at Great George Street. Look out for the Jubilee Walkway discs in the pavement as you go round. Directions. This walk starts from Trafalgar Square. Did you know? Trafalgar Square was laid out in 1840 by Sir Charles Barry, architect of the new Houses of Parliament. The square, which is now a 'World Square', is a place for national rejoicing, celebrations and demonstrations. It is dominated by Nelson's Column with the 18-foot statue of Lord Nelson standing on top of the 171-foot column. It was erected in honour of his victory at Trafalgar. With Trafalgar Square behind you and keeping Canada House on the right, cross Cockspur Street and keep right. Go around the corner, passing the Ugandan High Commission to enter The Mall under the large stone Admiralty Arch - go through the right arch. Keep on the right-hand side of the broad avenue that is The Mall. Did you know? Admiralty Arch is the gateway between The Mall, which extends southwest, and Trafalgar Square to the northeast. The Mall was laid out as an avenue between 1660-1662 as part of Charles II's scheme for St James's Park.
    [Show full text]
  • The Old War Office Building
    MINISTRY OF DEFENCE The Old War Office Building A history The Old War Office Building …a building full of history Foreword by the Rt. Hon Geoff Hoon MP, Secretary of State for Defence The Old War Office Building has been a Whitehall landmark for nearly a century. No-one can fail to be impressed by its imposing Edwardian Baroque exterior and splendidly restored rooms and stairways. With the long-overdue modernisation of the MOD Main Building, Defence Ministers and other members of the Defence Council – the Department’s senior committee – have moved temporarily to the Old War Office. To mark the occasion I have asked for this short booklet, describing the history of the Old War Office Building, to be published. The booklet also includes a brief history of the site on which the building now stands, and of other historic MOD headquarters buildings in Central London. People know about the work that our Armed Forces do around the world as a force for good. Less well known is the work that we do to preserve our heritage and to look after the historic buildings that we occupy. I hope that this publication will help to raise awareness of that. The Old War Office Building has had a fascinating past, as you will see. People working within its walls played a key role in two World Wars and in the Cold War that followed. The building is full of history. Lawrence of Arabia once worked here. I am now occupying the office which Churchill, Lloyd-George and Profumo once had.
    [Show full text]
  • View Our Entertaining Brochure [PDF 6.6MB]
    YOUR EVENT AT THE NATIONAL GALLERY Phone: +44 20 7747 5931 · Email: [email protected] · Twitter: @NG_EventsLondon · Instagram: @ngevents We are one of the most visited galleries in the world. Located on Trafalgar Square, we have stunning views of Nelson’s Column, Westminster and beyond. Surround your guests with some of the world’s greatest paintings by artists such as Leonardo, Monet and Van Gogh. Our central location, magnificent architecture and spectacular spaces never fail to impress. 2 The Portico The Mosaic Central Hall Room 30 The Barry Annenberg Gallery A The Yves Saint The Wohl Room The National Café The National Terrace Rooms Court Laurent Room Dining Room East Wing East Wing East Wing East Wing East Wing East Wing North Wing North Wing West Wing East Wing Sainsbury Wing Page 4 Page 5 Page 6 Page 7 Page 8 Page 9 Page 10 Page 11 Page 12 Page 14 Page 15 3 THE PORTICO With unrivalled views overlooking Trafalgar Square, the Portico Terrace is our most impressive outdoor event space. Breakfast 100 · Reception 100 · Reception including foyer 250 · Dinner 60 Level 2 DID YOU KNOW … The relief above the Portico entrance shows two young women symbolising Europe and Asia. This relief was originally intended for John Nash’s Marble Arch. 4 · East Wing THE MOSAIC TERRACE Step inside the Gallery onto our historic mosaic floor, beautifully lit by a domed glass ceiling. This unique artwork by Boris Anrep combines an ancient technique with contemporary personalities of 1930’s Britain such as Greta Garbo and Virginia Woolf.
    [Show full text]
  • Historical Portraits Book
    HH Beechwood is proud to be The National Cemetery of Canada and a National Historic Site Life Celebrations ♦ Memorial Services ♦ Funerals ♦ Catered Receptions ♦ Cremations ♦ Urn & Casket Burials ♦ Monuments Beechwood operates on a not-for-profit basis and is not publicly funded. It is unique within the Ottawa community. In choosing Beechwood, many people take comfort in knowing that all funds are used for the maintenance, en- hancement and preservation of this National Historic Site. www.beechwoodottawa.ca 2017- v6 Published by Beechwood, Funeral, Cemetery & Cremation Services Ottawa, ON For all information requests please contact Beechwood, Funeral, Cemetery and Cremation Services 280 Beechwood Avenue, Ottawa ON K1L8A6 24 HOUR ASSISTANCE 613-741-9530 • Toll Free 866-990-9530 • FAX 613-741-8584 [email protected] The contents of this book may be used with the written permission of Beechwood, Funeral, Cemetery & Cremation Services www.beechwoodottawa.ca Owned by The Beechwood Cemetery Foundation and operated by The Beechwood Cemetery Company eechwood, established in 1873, is recognized as one of the most beautiful and historic cemeteries in Canada. It is the final resting place for over 75,000 Canadians from all walks of life, including im- portant politicians such as Governor General Ramon Hnatyshyn and Prime Minister Sir Robert Bor- den, Canadian Forces Veterans, War Dead, RCMP members and everyday Canadian heroes: our families and our loved ones. In late 1980s, Beechwood began producing a small booklet containing brief profiles for several dozen of the more significant and well-known individuals buried here. Since then, the cemetery has grown in national significance and importance, first by becoming the home of the National Military Cemetery of the Canadian Forces in 2001, being recognized as a National Historic Site in 2002 and finally by becoming the home of the RCMP National Memorial Cemetery in 2004.
    [Show full text]
  • Northbank Book
    STRANDS OF HISTORY Northbank Revealed Clive Aslet Strands of History Northbank Revealed by Clive Aslet First published in 2014 by Wild Research, 40 Great Smith Street, London SW1P 3BU www.wildsearch.org © Wild Research 2014 All rights reserved The Northbank BID West Wing, Somerset House, Strand, London WC2R 1LA www.thenorthbank.org ISBN 978-0-9576966-2-4 Printed in Poland by ? ‘Looking to Northumberland House, and turning your back upon Trafalgar Square, the Strand is perhaps the finest street in Europe, blending the architecture of many periods; and its river ways are a peculiar feature and rich with associations.’ Benjamin Disraeli, Tancred: or, The New Crusade, 1847 ‘I often shed tears in the motley Strand for fullness of joy at so much life... Have I not enough, without your mountains?’ Charles Lamb, turning down an invitation from William Wordsworth to visit him in the Lake District Contents Foreword 10 Chapter One: The River 14 Chapter Two: The Road 26 Chapter Three: Somerset House 40 Chapter Four: Trafalgar Square 50 Chapter Five: Structural Strand: Charing Cross Station and Victoria Embankment 58 Chapter Six: Serious Strand: The Law Courts 64 Chapter Seven: Playful Strand: Shopping, Hotels and Theatres 72 Chapter Eight: Crown Imperial: The Strand Improvement Scheme 82 Chapter Nine: Art Deco and Post War 94 Chapter Ten: The Future 100 Image Acknowledgements 108 Further Reading 110 About Wild Research 111 7 8 About the Author Acknowledgements Clive Aslet is an award-winning writer and Maecenas molestie eros at tempor malesuada. journalist, acknowledged as a leading authority Donec eu urna urna.
    [Show full text]
  • The Architectural Transformation of Northumberland House Under the 7Th Duke of Somerset and the 1St Duke and Duchess of Northumberland, 1748–86
    THE ARCHITECTURAL TRANSFORMATION OF NORTHUMBERLAND HOUSE UNDER THE 7TH DUKE OF SOMERSET AND THE 1ST DUKE AND DUCHESS OF NORTHUMBERLAND, 1748–86 Adriano Aymonino and Manolo Guerci Adriano Aymonino, Department of Art History and Heritage Studies, University of Buckingham, Yeomanry House, Hunter Street, Buckingham MK18 1EG, UK. Email: adriano.aymonino@ buckingham.ac.uk Manolo Guerci, Kent School of Architecture, University of Kent, Marlowe Building, Canterbury CT27NR, UK. Email: [email protected] The material contained in this file consists of two appendices that should be read in con- junction with the paper published by Adriano Aymonino and Manolo Guerci in volume 96 of the Antiquaries Journal (2016) under the title ‘The architectural transformation of Northumberland House under the 7th Duke of Somerset and the 1st Duke and Duchess of Northumberland, 1748–86’. The first appendix is a list of the craftsmen and builders who worked on Northumber- land House during this period and the second is a transcription of an unpublished inventory made in 1786 at the death of Sir Hugh Smithson (1712–86), 2nd Earl and later 1st Duke of Northumberland. Works referred to in the footnotes are listed in the bibliography at the end of this file, which also lists the published and unpublished sources referred to in the main paper, which can be found on Cambridge University Press’s online publishing platform, Cambridge Core: cambridge.org/core/; doi: 10.1017/s0003581516000676 APPENDIX 1 Craftsmen and builders employed at Northumberland House, 1748–58 Information on the craftsmen and builders employed at Northumberland House derives mostly from the detailed accounts dating from 1748 to 1767 of the 7th Duke of Somerset and Lord Northumberland with ‘Messrs.
    [Show full text]