Default view
Week
Month
Name
Link
Job
Profile
Content
Date
Likes
Comments
ML Score
Image Link
Jonathan Regenstein
Open
https://www.linkedin.com/feed/update/urn:li:activity:6933777022309588992
Truist Securities | Head of Data and Quantamental Research
https://www.linkedin.com/in/jkregenstein/
What a difference 40 years makes....Sudheer Chava and I are wrapping up our inflation chapter with a look at market conditions in different inflationary environments. In 2022 we find ourselves printing CPI YoY changes consistently over 7%. As of April, before the recent, massive market selloff, the SP500's cyclically adjusted P/E ratio (the CAPE ratio) was above 34. That's a high CAPE ratio in any environment, but in a 7% inflationary environment, it's outlier status jumps out of the boxplot (link to a great write up by Lyn Alden Schwartzer on the CAPE in the comments below).We also label CAPE ratios below 7 and this takes us back to 1982. The economy was mired in a recession at this time, Volker had raised interest rates to crush inflation and unemployment was high. Needless to say, today's CAPE could leg down significantly in the coming months. Good luck out there!#rstats #macroeconomics
2022/05/21
44
1
3
https://media-exp1.licdn.com/dms/image/C5622AQGHpSlJRNf8wA/feedshare-shrink_800/0/1653141264482?e=1655942400&v=beta&t=nEyrX3E-Xsn-g-kuKcq0naKB1yfmvR4S6ESbM60sNwI
Charles-Albert Lehalle
Open
https://www.linkedin.com/feed/update/urn:li:activity:6933433694275653634
Abu Dhabi Investment Authority (ADIA) | Quantitative R&D Lead
https://www.linkedin.com/in/lehalle/
venez échanger le mardi 31 mai à Paris sur les stratégies d'investissement systématiques avec Marcos Lopez de Prado, Gautier Marti, Marie Brière, Hugues Langlois, Jeroen VK Rombouts et votre serviteur à ce workshop organisé par l'ILB. Au programme: Open Problems in Quantitative Asset Management, Alternative Data et Machine Learning.
2022/05/20
87
5
9
Petter Kolm
Open
https://www.linkedin.com/feed/update/urn:li:activity:6933461572040232961
Courant Institute of Mathematical Sciences | Clinical Full Professor of Mathematics
https://www.linkedin.com/in/petterkolm/
I am looking forward to presenting at the Machine Learning and Quantitative Finance Workshop on June 1 at the Oxford-Man Institute of Quantitative Finance, University of Oxford. More info here:https://lnkd.in/gr9NhMBk#nyu #nyucourant #machinelearning #financialdatascience #trading
2022/05/20
55
0
5
https://media-exp1.licdn.com/dms/image/C5622AQGtdu5EGpAhuQ/feedshare-shrink_800/0/1653066054573?e=1655942400&v=beta&t=MPhEqtzF1iytnxDZ5BQNExO6sasiYLzX58c420YcJSI
Graham Giller
Open
https://www.linkedin.com/feed/update/urn:li:activity:6933217486058250240
Chief Executive Officer |
https://www.linkedin.com/in/grahamgiller/
Is it wrong to be so happy about this? Leslie Lamport, inventor of LaTeX, recently gave an interview in which he hoped that computer scientists would think more about mathematics before writing code. (https://lnkd.in/gRsPmimV)Well ... now they can communicate those thoughts more easily because after 8 years of waiting, GitHub has introduced LaTeX rendering! All things come to those who wait. #mathematics #latex #github #datascience
2022/05/20
646
27
3
https://media-exp1.licdn.com/dms/image/sync/C4D27AQH6G4BXcatPsg/articleshare-shrink_800/0/1653058150492?e=1653685200&v=beta&t=JGvgKmev4sRMhuY5gu-4OqdXo475RzaBcLjHodmga-s
Peter Cotton, PhD
Open
https://www.linkedin.com/feed/update/urn:li:activity:6933488983138082816
Intech Investment Management LLC | Senior Vice President. Chief Data Scientist.
https://www.linkedin.com/in/petercotton/
For those interested in portfolio optimization, here is the list of all Riskfolio-lib (1186 
️ on Github) examples with the name of the example (https://lnkd.in/d2vdGJZi). Thanks Peter Cotton, PhD for made this list.#investments #portfoliooptimization #quant #finance #assetmanagement #riskmanagement #assetmanagement
️ on Github) examples with the name of the example (https://lnkd.in/d2vdGJZi). Thanks Peter Cotton, PhD for made this list.#investments #portfoliooptimization #quant #finance #assetmanagement #riskmanagement #assetmanagement2022/05/20
52
3
3
https://media-exp1.licdn.com/dms/image/sync/C4E27AQHk_KvToTRE2Q/articleshare-shrink_1280_800/0/1653071235426?e=1653685200&v=beta&t=VdVCBiey3jcOa3244D7uAZ3VSJ2wZgzLUpvw1phy-rI
Igor Halperin
Open
https://www.linkedin.com/feed/update/urn:li:activity:6933383758725218304
Fidelity Investments | AI Asset Management
https://www.linkedin.com/in/igor-halperin-092175a/
“I was reading the number of the bus when the bus was arriving.” With a curiosity that can be compared to a child’s, physics laureate Giorgio Parisi tells us about how he started to read numbers at the age of 3. At an early age, he also started to read the Russian author Leo Tolstoy as his parents’ large library encouraged reading. In our newest podcast episode we follow Parisi’s life journey and how he ultimately decided to dedicate his life to physics as that topic made more sense to him than mathematics. In 2021 he was awarded the physics prize for his work on complex systems. Listen here: https://lnkd.in/e9K-bmZp
2022/05/20
1623
10
2
https://media-exp1.licdn.com/dms/image/C4E22AQH_llyQ_MJKdA/feedshare-shrink_800/0/1652983021784?e=1655942400&v=beta&t=CO5PwKwVVAgzF_lMy9RqkSwZxQfhmRv6aGf5XcBflLU
Ariel Silahian
Open
https://www.linkedin.com/feed/update/urn:li:activity:6932817352912769024
Quant Developers | High Frequency Trading solutions | Founder & CTO / Lead Tech
https://www.linkedin.com/in/silahian/
!
2022/05/19
17
9
4
https://media-exp1.licdn.com/dms/image/C5622AQHsqSrbNoFQAg/feedshare-shrink_800/0/1652906577690?e=1655942400&v=beta&t=cRD-wQxFG_ANMDVY11KaV0SMqKp_ZQuxm8K8IftbGs8
Spyros Makridakis
Open
https://www.linkedin.com/feed/update/urn:li:activity:6933055251461668864
University of Nicosia | Professor
https://www.linkedin.com/in/spyros-makridakis-b2ba5a52/
In about 10 days the first quarter of the M6 competition is ending with the winners being awarded $42,000 in prize money. It is interesting the top four teams in the Global score are two experienced data scientists and two students, It seems that the pattern of the M5 competition is followed with a student winning the accuracy challenge and a team of two data scientists winning the uncertainty one, May be the time has come that students can master ML methods equally well than experienced data scientists! Maybe AI is bringing some big changes in the value of experience and expertise in all aspects of our work and lives.
2022/05/19
35
4
3
https://media-exp1.licdn.com/dms/image/C5622AQFNt4HrhAFUZQ/feedshare-shrink_2048_1536/0/1652969180965?e=1655942400&v=beta&t=UFZujYcQTobmSgk5pVTXQgAuTDYAWzS-ru05wkjZkfc
Peter Cotton, PhD
Open
https://www.linkedin.com/feed/update/urn:li:activity:6933148542165274625
Intech Investment Management LLC | Senior Vice President. Chief Data Scientist.
https://www.linkedin.com/in/petercotton/
One of the 2022 gold papers for the new AI world! - Model-centric AI to Data-centric AI. This is all because it seems like there is an AI predictive performance limit given a data set. So, if you want to further improve prediction, attend to the data, not keep grinding on the modeling methodology. Super excited about this Innovative data solution for the hungry AI. Data engineering will have to include these ideas in curriculum development soon. The video is a year old...Keep on trying to improve the model does not help because the data is not right for the solution. A powerful collection of case studies are presented in the video. This is the height of new world data engineering! AI sees differently, reads differently, hears differently, and senses differently. Do not feed more and more for the hungry AI. Develop and feed data smartly, and we may create AI solutions with smaller data. A plan on how to leverage different aspects of collection, augmentation, distortion, simulation, and GANizing data for failure sample points can reduce the data needs and hence overall smaller data sets!! This has a tinge of boosting in algorithm building, but for adding data points, boosting the data collection for AI performance, rather than boosting, boosting,..., and adding more and more hidden layers in the analytics. "It is useful to shift our mindset improving data systematically" - https://lnkd.in/g43MwyQf has the video version of this topic. He explains also the importance of DevOps and why a traditional software engineer does not seem to be popular in this team, but an ML Engineer is in the eyes of the managers!The interesting question is, "is data science going more and more on designed data", which is what statisticians have been doing for eons. Thank you, Prof. Andrew Ng #artificialintelligence #machinelearning #datasciencetraining #datasciences #datasc
AI. Data engineering will have to include these ideas in curriculum development soon. The video is a year old...Keep on trying to improve the model does not help because the data is not right for the solution. A powerful collection of case studies are presented in the video. This is the height of new world data engineering! AI sees differently, reads differently, hears differently, and senses differently. Do not feed more and more for the hungry AI. Develop and feed data smartly, and we may create AI solutions with smaller data. A plan on how to leverage different aspects of collection, augmentation, distortion, simulation, and GANizing data for failure sample points can reduce the data needs and hence overall smaller data sets!! This has a tinge of boosting in algorithm building, but for adding data points, boosting the data collection for AI performance, rather than boosting, boosting,..., and adding more and more hidden layers in the analytics. "It is useful to shift our mindset improving data systematically" - https://lnkd.in/g43MwyQf has the video version of this topic. He explains also the importance of DevOps and why a traditional software engineer does not seem to be popular in this team, but an ML Engineer is in the eyes of the managers!The interesting question is, "is data science going more and more on designed data", which is what statisticians have been doing for eons. Thank you, Prof. Andrew Ng #artificialintelligence #machinelearning #datasciencetraining #datasciences #datasc2022/05/19
137
19
1
https://media-exp1.licdn.com/dms/image/sync/C5627AQEa5558TtNquA/articleshare-shrink_800/0/1652962354127?e=1653598800&v=beta&t=klhoDgvtpaU9ZBe30R63ovXocg-iZwQa1eIdx6halYo
Prof. Alexander Lipton
Open
https://www.linkedin.com/feed/update/urn:li:activity:6933037395143458816
Sila | Co-Founder, Chief Information Officer
https://www.linkedin.com/in/prof-alexander-lipton-aa2256bb/
A powerful keynote from Edward Jung on how to solve the world's biggest challenges of tomorrow with innovation. Few are better placed to speak about innovation than him.At an event addressing Sheikh Mohamed bin Zayed’s (UAE’s new President) council gathering over Ramadan, Edward proposed that global innovation promoting collaboration and integration between policy and technology is the only way to solve big global problems.Small, agile countries have big advantages over large countries in this new innovation economy.As our friends know, the principles of collaboration, teamwork, and agility are core values for us at Katalysen Ventures, and we like Edward believe in the Power of Small. Edward is a Katalysen adviser and investor.Interested to learn more about how we make the Power of Small work in venture development? Read more here: https://lnkd.in/eAcNs6KQWatch the keynote: https://lnkd.in/egNtHxrA
2022/05/19
4
2
1
Jacques Joubert
Open
https://www.linkedin.com/feed/update/urn:li:activity:6932772811241316352
Shell | Systematic Trader
https://www.linkedin.com/in/jacquesjoubert/
*** Market Microstructure - Quantitative Research ***I am happy to share my last paper on market impact which completes the previous ones that I published on equity and options market during the last years (links in comments). This third paper completes the two previous ones and closes this market impact trilogy I started 6 years ago. In this paper, we propose a theory of the market impact of metaorders based on a coarse-grained approach where the microscopic details of supply and demand is replaced by a single parameter ρ ∈ [0, +∞] shaping the supply-demand equilibrium and the market impact process during the execution of the metaorder. Our model provides an unified explanation of most of the empirical observations that have been reported and establishes a strong connection between the excess volatility puzzle and the order-driven view of the markets through the square-root law.I would like to thank Marcos Lopez de Prado and Prof. Alexander Lipton for their comments on the preliminary version of this paper. I am also particularly grateful to Charles-Albert Lehalle for his careful reading, comments and the many interesting discussions we had about it.
2022/05/18
0
0
10
Harvey Stein
Open
https://www.linkedin.com/feed/update/urn:li:activity:6932706402670960640
Bloomberg LP | Head, Quantitative Risk Analytics
https://www.linkedin.com/in/harveyjstein/
Interested in draw-downs? Stan Uryasev and Rui D. will be presenting their research on draw-down based risk measures and portfolio optimization with respect to them in the next to last IAQF/Thalesians webinar of the semester. Monday, May 23rd at 12:30 NY time.#research #drawdown #riskmanagement #riskanalytics #quantitativefinance #webinar #portfoliooptimization #optimization
2022/05/18
27
0
3
https://media-exp1.licdn.com/dms/image/sync/C4E27AQE-6K5ji0ek3w/articleshare-shrink_1280_800/0/1652886043898?e=1653512400&v=beta&t=Hb8yfLdciQLfnL02z7xj6uiLjKS1q-mQeYwmI9POcM0
Derek Snow
Open
https://www.linkedin.com/feed/update/urn:li:activity:6932688374638002176
Oxford-Man Institute of Quantitative Finance, University of Oxford | Associate Member
https://www.linkedin.com/in/snowderek/
Explainable and interpretable models in machine learning are not easily distinguished. Interpretable models are convenient because they show the exact contribution of each feature to the final result.With non-interpretable models, you only get an estimate of the contribution of each feature to the final outcome. Explainable models then seem to bridge the difference between these models by using explanatory methods, some of which are listed below.We can say that non-interpretable models are blackbox models, interpretable models are whitebox models, and non-interpretable models + explainability techniques are greybox models. I have been thinking about how to represent this idea. notion page: https://lnkd.in/gjJ2-Mg6 #quantitativeresearch #algotrading #explainability
2022/05/18
21
2
3
https://media-exp1.licdn.com/dms/image/C5622AQEYt6bhF8otHg/feedshare-shrink_800/0/1652881710927?e=1655942400&v=beta&t=pe5KNd5z2Qkhmu2_TxcBR0M4aXhWpWp8KNYDUjML-fs
Wesley Gray, Ph.D.
Open
https://www.linkedin.com/feed/update/urn:li:activity:6932683043845402624
Alpha Architect, LLC | CEO
https://www.linkedin.com/in/alphaarchitect/
Plenty of research shows that #insider buys contain value-relevant information while insider sales include little to no information. But what about the action of “not trading”?Are the trades of portfolio insiders informative about the stocks they choose not to trade, those they choose not to buy, and those they choose not to sell?To finish reading, article by Elisabetta B. is in comments field.
But what about the action of “not trading”?Are the trades of portfolio insiders informative about the stocks they choose not to trade, those they choose not to buy, and those they choose not to sell?To finish reading, article by Elisabetta B. is in comments field.2022/05/18
10
4
3
https://media-exp1.licdn.com/dms/image/C5622AQHaeTfxlG5kYg/feedshare-shrink_2048_1536/0/1652880439977?e=1655942400&v=beta&t=LqGq3Y0VnltoxgNWcmqCGEbkKwf5LpJyMvzoBFrU0tk
Spyros Makridakis
Open
https://www.linkedin.com/feed/update/urn:li:activity:6932715558685569024
University of Nicosia | Professor
https://www.linkedin.com/in/spyros-makridakis-b2ba5a52/
Peter before going after Crazy.Jim you need first to beat me. It cannot be difficult!
2022/05/18
0
0
3
https://media-exp1.licdn.com/dms/image/C4E22AQE41u_-RlG2Vg/feedshare-shrink_800/0/1652883100825?e=1655942400&v=beta&t=RclYVudqjLNqpvugkp_wEUpbUVyQfiXN8XnVhfFSSC0
Prof. Alexander Lipton
Open
https://www.linkedin.com/feed/update/urn:li:activity:6932647877668130816
Sila | Co-Founder, Chief Information Officer
https://www.linkedin.com/in/prof-alexander-lipton-aa2256bb/
I solved my biggest crypto trading problem.I didn't have any time and was always stressed.This is how I did it:I built an algorithm that learns to trade by itself.It starts by developing two different, random trading strategies.And then compares how they perform on a set of data.The best strategy is then selected and a new strategy is created, and they are backtested again.The algorithm does this thousands of times over, until the portfolio's requirements are met.This strategy has done 3 things for me:1. Reduced my portfolio risk2. Increased my returns3. Reduced how much time I spend trading#crypto #bitcoin #ethereum #blockchain #trading
2022/05/18
203
74
2
https://media-exp1.licdn.com/dms/image/C4D22AQFv-YTXM5J40Q/feedshare-shrink_2048_1536/0/1652861774635?e=1655942400&v=beta&t=jPSPyk8OFoKWy3KPAhZzkoz8DjpOaWf4-dCcv8y9H3w
Harvey Stein
Open
https://www.linkedin.com/feed/update/urn:li:activity:6932077933444407296
Bloomberg LP | Head, Quantitative Risk Analytics
https://www.linkedin.com/in/harveyjstein/
Our white paper on default modeling is on https://lnkd.in/eUMAPYHU. Nick Costanzino, Albert Cohen and I found that looking at stopping times in terms of their conditional survival curves yields a general framework for default modeling that encompasses both structural and reduced form models. Moreover, whether the stopping time is predictable or totally inaccessible can be directly read off of the survival curves.#default #creditrisk #stoppingtime #predictable #inaccessible #whitepaper
2022/05/17
75
1
5
https://media-exp1.licdn.com/dms/image/sync/C4E27AQGdlaNFuAfbPA/articleshare-shrink_480/0/1652735941669?e=1653426000&v=beta&t=rVTsVL_ZTtDW19laMG9Pnua-Ubr56hz1cPck7XM7XI8
Jacques Joubert
Open
https://www.linkedin.com/feed/update/urn:li:activity:6932378149317939200
Shell | Systematic Trader
https://www.linkedin.com/in/jacquesjoubert/
ML-Quant now has 6 additional web scrapers, for a total of 30 scrapers working 24/7 to 'find' trending content in the quant finance and machine learning space. Today, I added a script to find the top general ML papers as per their social media metrics (twitter, news, YouTube etc.) https://www.ml-quant.com/
2022/05/17
3
0
5
https://media-exp1.licdn.com/dms/image/C5622AQExVdUYpFdggw/feedshare-shrink_2048_1536/0/1652796802965?e=1655942400&v=beta&t=s3QHVj6ve8mZrPLBj6wOXcJ8ePjoz0n1pPwzEnI6ncg
Dr Miquel Noguer i Alonso
Open
https://www.linkedin.com/feed/update/urn:li:activity:6932364172722655233
Artificial Intelligence Finance Institute - AIFI | Founder at Artificial Intelligence Finance Institute
https://www.linkedin.com/in/dr-miquel-noguer-i-alonso-7242345/
We released a new paper : " Deep Signature models for Financial Equity Time Series prediction" along with Sonam Srivastava and Himanshu AgrawalWe explore in this paper the use of deep signature models to predict equity financial time series returns. First, we use signature transformations to model the underlying shape of the input equity returns; further assuming the underlying shape remains the same, we predict future values based on that shape. Finally, different neural networks are used to process the output from signature transformation to predict equity returns: Long Short Term Memory Networks, Signet Model, and Deep Signature Model. Feeding signature transformations to a neural network brings significant improvement in prediction. Using signature transformation and Long Short Term Memory Networks proves to be the best performing model in accuracy and precision. In contrast, on RMSE terms, all three models offer very comparable performance.You can download the paper here:https://lnkd.in/gWqHeBn5www.aifinanceinstitute.com AIFI - Artificial Intelligence Finance Institute
2022/05/17
50
0
4
https://media-exp1.licdn.com/dms/image/C5622AQEtDAV4zXJEAw/feedshare-shrink_800/0/1652804414293?e=1655942400&v=beta&t=LRM-9OsZrSppsqLzUamU733hL4A33vyuh7Db-fVdKUY
Álvaro Cartea
Open
https://www.linkedin.com/feed/update/urn:li:activity:6932267187738951680
University of Oxford | Professor
https://www.linkedin.com/in/alvaro-cartea/
Conference on Natural Language Processing for Economic and Financial Modelling, University of Oxford. 22-23 June #nlp #machinelearning
2022/05/17
31
1
3
https://media-exp1.licdn.com/dms/image/sync/C4E27AQFaP5jdsaxuwA/articleshare-shrink_480/0/1652779930923?e=1653426000&v=beta&t=sFRZS562_DWYJEiALQHAE0xIuLfrbrPXL3DiD7PLdVA
Jochen Papenbrock
Open
https://www.linkedin.com/feed/update/urn:li:activity:6932312605554311168
NVIDIA | Financial Services and Technology Developer Relationship Lead EMEA
https://www.linkedin.com/in/jochenpapenbrock/
At the STAC Summit on May 19th, NVIDIA's Tim Wood is speaking about scalable platforms for HPC and innovation in AI. Register now.
2022/05/17
1
0
2
https://media-exp1.licdn.com/dms/image/sync/C4E34AQHmBp4Ntfn_jw/ugc-proxy-shrink_800/0/1652792121872?e=1653426000&v=beta&t=CIR5Po78oyOcG0_Q6nuMCKLuIgR3VsQHCOPcUguf_7M
Jochen Papenbrock
Open
https://www.linkedin.com/feed/update/urn:li:activity:6932407186161598464
NVIDIA | Financial Services and Technology Developer Relationship Lead EMEA
https://www.linkedin.com/in/jochenpapenbrock/
This NVIDIA Deep Learning tutorial at #ISC22 introduces the NVIDIA solution for quantum-classical computing. See a hands-on tutorial demonstrating how #GPU-based quantum-classical programming is accelerating scientific computing workflows. Register now.
2022/05/17
0
0
2
https://media-exp1.licdn.com/dms/image/sync/C4E34AQFxjHprF4IAUw/ugc-proxy-shrink_800/0/1652814672139?e=1653426000&v=beta&t=3S1nWGpmh1wKeNZpE3pMkvQZJcqUUIe_OZX7DlNilJA
Jochen Papenbrock
Open
https://www.linkedin.com/feed/update/urn:li:activity:6932395452784893954
NVIDIA | Financial Services and Technology Developer Relationship Lead EMEA
https://www.linkedin.com/in/jochenpapenbrock/
Discussing CUDA-accelerated data science with John Murray and Daniel Jünger in the Eisgrubbräu brewery #nvidia
#nvidia2022/05/17
23
3
1
https://media-exp1.licdn.com/dms/image/C4E22AQFZJ4vsD2osGg/feedshare-shrink_800/0/1652810065075?e=1655942400&v=beta&t=MSRp7h3iF5IDDI9bvjOeZAp7ZEOG_straQJgwyYpU6M
Álvaro Cartea
Open
https://www.linkedin.com/feed/update/urn:li:activity:6931961251757490176
University of Oxford | Professor
https://www.linkedin.com/in/alvaro-cartea/
New white paper out with my co-authors at the University of Oxford.OTC market making, Artificial intelligence, independent multi-agent learning, risk of pseudo collusion, and much more!Hope you enjoy it - feedback most welcome.
2022/05/16
8
0
3
Derek Snow
Open
https://www.linkedin.com/feed/update/urn:li:activity:6932006615067852800
Oxford-Man Institute of Quantitative Finance, University of Oxford | Associate Member
https://www.linkedin.com/in/snowderek/
I am logging content produced in the Quant Finance/Machine Learning space. It's probably a fool's errand, but I sometimes want to take a break from our 'adaptive' trade. I will post a few versions over the next few weeks. Please let me know if you have any recommendations, its easy to change and rerun the script. #machinelearning #quantitativeresearch #algotrading #finance
2022/05/16
29
3
2
Peter Cotton, PhD
Open
https://www.linkedin.com/feed/update/urn:li:activity:6931998226317234176
Intech Investment Management LLC | Senior Vice President. Chief Data Scientist.
https://www.linkedin.com/in/petercotton/
Machine Learning for High-frequency Time SeriesGlobal parametric functional models (like neural networks for example) are not particularly suitable for short-term prediction of high-frequency time series of prices/returns. The theoretical reason is that such financial series are typically aperiodic with non-Gaussian noise. These characteristics require large number of nonlinear terms when building global models [Box and Jenkins, 1974] in order to achieve parameter stability and adequate fitting (over the entire training temporal interval). Larger models, however, tend to overfit [Geman et al.,1992] and produce poor short-term forecasts (usually with time delay). This is a difficult dilemma which is practically impossible to resolve. The proper approach to learning from stochastic financial series for algorithmic trading is to use nonparametric local models [Farmer and Sidorowich, 1998]. The plot below shows rolling one-step ahead predictions by a local model (of AAPL mid-prices recorded at 30-seconds on 2/1/2020) without phase delay which directional accuracy is 68.53% hits.
2022/05/16
41
11
1
https://media-exp1.licdn.com/dms/image/C4E22AQGgRQVRhnjKDg/feedshare-shrink_800/0/1652703913761?e=1655337600&v=beta&t=GV5YfEzyKuC9v7P1M1U2lFD2PvmTmVbc5HZrPoJyEnk
Charles-Albert Lehalle
Open
https://www.linkedin.com/feed/update/urn:li:activity:6931464421030400000
Abu Dhabi Investment Authority (ADIA) | Quantitative R&D Lead
https://www.linkedin.com/in/lehalle/
Researchers in speech recognition, computer vision, and natural language processing in the 2000s were obsessed with accurate representations of uncertainty. This led to a flurry of work on probabilstic generative models such as Hidden Markov Models in speech, Markov random fields and constellation models in vision, and probabilistic topic models in NLP, e.g. with latent Dirichlet analysis.There were debates at computer vision workshops about "generative models vs discriminative models". There were heroic-yet-futile attempts to build object recognition systems with non-parametric Bayesian methods.Much of this was riding on previous work on Bayesian networks, factor graphs and other graphical models. That's how one learned about exponential family, belief propagation, loopy belief propagation, variational inference, etc, Chinese restaurant process, Indian buffet process, etc.But almost none of this work was concenred with the problem of learning representations. Features were assumed to be given. The structure of the graphical model, with its latent variables, was assumed to be given.All one had to do was to compute some sort of log likelihood by linearly combining features, and then use one of the above-mentioned sophisticated inference methods to produce marginal distributions over the unknown variables, one of which being the answer, e.g. a category.In fact, "exponential family" pretty much means "shallow": the log-likelihood can be expressed as a linearly parameterized function of features (or simple combinations thereof).Learning the parameters of the model was seen as just another variational inference problem.It's interesting to observe that almost none of this is relevant to today's top speech, vision, and NLP systems.As it turned out, solving the problem of learning hierarchical representations and complex functional dependencies was a much more important issue
2022/05/15
4097
132
3
https://media-exp1.licdn.com/dms/image/sync/C5627AQHsocwDKU5W6w/articleshare-shrink_800/0/1652561970898?e=1653253200&v=beta&t=Rv9YEcMx4A5OcPHCJeshHDS0B8eXRKVPJxJOzfehH90
Spyros Makridakis
Open
https://www.linkedin.com/feed/update/urn:li:activity:6931537778585317376
University of Nicosia | Professor
https://www.linkedin.com/in/spyros-makridakis-b2ba5a52/
Below is the graph of the RPS (Ranked Probability Scores) for all teams. The graph shows that only 22% of teams beat the benchmark (versus 60% that did so in IR, Information Ratio). On the negative side, the 22% is very low and disappointing. On the positive side, it implies a great opportunity for learning and achieving considerable improvements. In our view, the battle for winning the M6 will be won by achieving substantial improvements in the RPS part that will definitely be more challenging than the IR ones but counting equally to the overall rank. Improving the RPS has not received much attention in the academic literature and is the essence of the M6 to learn as much as possible and advance its understanding and accuracy. We have high hopes that this will be achieved and that the M6 will allow us to explore the link between forecasting accuracy and investment returns, hoping to achieve substantial improvements in the RPS part.Evangelos Spiliotis
2022/05/15
30
1
2
https://media-exp1.licdn.com/dms/image/C4E22AQGLOest6qdf0g/feedshare-shrink_800/0/1652607387584?e=1655337600&v=beta&t=WejG5E0ijZnUj8zLhsozdh8H95suWYtTNk_3Yw3oX6c
Charles-Albert Lehalle
Open
https://www.linkedin.com/feed/update/urn:li:activity:6931130718018252802
Abu Dhabi Investment Authority (ADIA) | Quantitative R&D Lead
https://www.linkedin.com/in/lehalle/
# AAD #LocalVol #implementation #Collective #ComputationTime #Accuracy #Speed
2022/05/14
88
15
2
https://media-exp1.licdn.com/dms/image/C4E12AQFC5zpK91F9cA/article-cover_image-shrink_423_752/0/1652419947259?e=1657756800&v=beta&t=Cf22GI9p999ru0usVVIbXiIGXMkZuWoBwSldg8ie3Bo
Peter Cotton, PhD
Open
https://www.linkedin.com/feed/update/urn:li:activity:6931226666836893696
Intech Investment Management LLC | Senior Vice President. Chief Data Scientist.
https://www.linkedin.com/in/petercotton/
Interesting reading for those who suspect the future of AI might be coordinated by billions of little self-interested buzzards (not by powerpoint slides) ... at least in the domain where automatic assessment is possible. If we define a micromanager as an autonomous reward seeking prediction agent that establishes and maintains small economic relationships with others of its ilk, in order to improve its own predictions upstream, then we see it can take many forms. The space of generalized cost-aware regression mechanisms with accompanying incentive schemes is vast, for instance. So too the set of market-inspired mechanisms, and many constructs half-way in between (market scoring rules and the like). It seems the energy forecasting literature is a rich source of inspiration, as you can read here. #machinelearning #automl #microprediction
2022/05/14
33
3
2
https://media-exp1.licdn.com/dms/image/sync/C4E27AQEOaqklIWOlqw/articleshare-shrink_1280_800/0/1652250543002?e=1653166800&v=beta&t=yjvhtRtYQZ3EK9RvGDIYqKIcqMQvMfBGjU31cHrqBHU
Prof. Alexander Lipton
Open
https://www.linkedin.com/feed/update/urn:li:activity:6930869869294804992
Sila | Co-Founder, Chief Information Officer
https://www.linkedin.com/in/prof-alexander-lipton-aa2256bb/
Continuing my previous post, I'm happy to present the latest episode of Quantcast with the quant finance editor at Risk.net, Mauro Cesa, https://lnkd.in/dYp4hA2r). In addition to stablecoins, we also discussed my recent paper with Artur Sepp of Sygnum Bank on automated market-making in FX, https://lnkd.in/ddVw7RnR. It describes "how central bank digital currencies or stablecoins can be exchanged through a smart contract on the blockchain while retaining pricing consistent with a traditional centralized market." I also argue that "the approach would make FX markets more transparent, retaining the monetary incentives for market-makers while improving efficiency for other players." It also allows direct exchange of relatively illiquid currencies without the need for US dollar transactions.Index: 00:00 Introduction 05:00 Automated FX market-making 09:10 On-chain and off-chain interaction 11:25 Applications of the framework 12:25 The problems with algorithmic stablecoins 20:00 The collapse of TerraUSD 22:55 Viable blockchain applications in finance 29:30 The limits of DeFi 34:35 Non-financial applications#blockchain #stablecoins #fx
2022/05/13
42
2
8
https://media-exp1.licdn.com/dms/image/sync/C4E27AQFx3SFb6ahwBA/articleshare-shrink_480/0/1652448160028?e=1653080400&v=beta&t=l2DfTFzlmFDy7pQdOb24KwrMGLcU6hNG9xn18AGjL7s
Igor Halperin
Open
https://www.linkedin.com/feed/update/urn:li:activity:6930860574595571712
Fidelity Investments | AI Asset Management
https://www.linkedin.com/in/igor-halperin-092175a/
Tuesday, May 24, 5:30PM-6:45PM (Eastern Time): Join us for Igor Halperin's online talk "Combining Reinforcement Learning and Inverse Reinforcement Learning for Asset Allocation Recommendations" in NYU Courant's Mathematical Finance & Financial Data Science Seminar. This online event is open to the public, but requires registration. Registration & more details:https://lnkd.in/dzF832kmAbstract: We suggest a simple practical method to combine the human and artificial intelligence to both learn best investment practices of fund managers, and provide recommendations to improve them. Our approach is based on a combination of Inverse Reinforcement Learning (IRL) and RL. First, the IRL component learns the intent of fund managers as suggested by their trading history, and recovers their implied reward function. At the second step, this reward function is used by a direct RL algorithm to optimize asset allocation decisions. We show that our method is able to improve over the performance of individual fund managers.#nyucourant #nyu New York University Courant Institute of Mathematical Sciences NYU Courant Institute of Mathematical Sciences M.S. in Mathematics in Finance, NYU Courant New York University #reinforcementlearning #machinelearning #trading #inversereinforcementlearning
2022/05/13
86
8
6
https://media-exp1.licdn.com/dms/image/C4D22AQEDTgGRD03HOg/feedshare-shrink_800/0/1652390477638?e=1655337600&v=beta&t=NsGGl0MBwp3Z8b1nH0U2aATti7BCfs55G5kuYPEP1Zs
Mauro Cesa
Open
https://www.linkedin.com/feed/update/urn:li:activity:6930818141681819648
Infopro Digital | Quant finance editor, Risk.net
https://www.linkedin.com/in/mauro-cesa/
For the latest Quantcast, I’m joined by Prof. Alexander Lipton, global head of quantitative R&D at ADIA. He talks about his latest work with Artur Sepp on their automated market making framework for currencies that uses smart contracts and central bank digital currencies or stablecoins. From stablecoins, the conversation inevitably moves to the flaws of algorithmically stabilised coins like Terra UST, and why their design cannot work.
2022/05/13
35
2
4
https://media-exp1.licdn.com/dms/image/sync/C4E27AQFx3SFb6ahwBA/articleshare-shrink_480/0/1652448160028?e=1653080400&v=beta&t=l2DfTFzlmFDy7pQdOb24KwrMGLcU6hNG9xn18AGjL7s
Peter Cotton, PhD
Open
https://www.linkedin.com/feed/update/urn:li:activity:6930662951401975808
Intech Investment Management LLC | Senior Vice President. Chief Data Scientist.
https://www.linkedin.com/in/petercotton/
I encourage all of you to attend The Peter Carr Memorial Conference to honor the life and career of our beloved friend and colleague. Peter we will forever remember you, your wit, your smarts and your friendship. You were one of a kind. #nyu #nyucourant #nyutandon #petercarr Dilip Madan
2022/05/13
34
0
3
https://media-exp1.licdn.com/dms/image/C4E22AQE9_ytuMfTynw/feedshare-shrink_800/0/1652380231011?e=1655337600&v=beta&t=IY1Q347KEVSymW8e3ho5_Mw6PloETspyotPC2L4pmJk
Thierry Roncalli
Open
https://www.linkedin.com/feed/update/urn:li:activity:6930754560617549824
Amundi | Head of Quantitative Research
https://www.linkedin.com/in/thierry-roncalli-78a98b12b/
Wednesday, May 18, 5:30PM-6:45PM (Eastern Time): Join us for Thomas Huckle and Jonathan Raimana Chan's online talk "Portfolio Optimisation with Options" in NYU Courant's Mathematical Finance & Financial Data Science Seminar. This online event is open to the public, but requires registration. Registration & more details:https://lnkd.in/dqZWuyQNAbstract: We develop a new analysis for portfolio optimisation with options, tackling the three fundamental issues with this problem: asymmetric options' distributions, high dimensionality and dependence structure. To do so, we propose a new dependency matrix, built upon conditional probabilities between options' payoffs, and show how it can be computed in closed form given a copula structure of the underlying asset prices. The empirical evidence we provide highlights that this approach is efficient, fast and easily scalable to large portfolios of (mixed) options. This is join work with Jack Jacquier and Aitor Muguruza Gonzalez.#nyucourant #nyu New York University Courant Institute of Mathematical Sciences NYU Courant Institute of Mathematical Sciences M.S. in Mathematics in Finance, NYU Courant New York University #portfoliomanagement #portfoliooptimization #trading #options
2022/05/13
28
0
3
https://media-exp1.licdn.com/dms/image/C4D22AQHLXxuOThxTQA/feedshare-shrink_800/0/1652389951373?e=1655337600&v=beta&t=vx1TkhkZbSNkwKvYJgFtVWnItavkXT4DQjQY1Yc4Sbk
Álvaro Cartea
Open
https://www.linkedin.com/feed/update/urn:li:activity:6930801746604658688
University of Oxford | Professor
https://www.linkedin.com/in/alvaro-cartea/
New paper with Tao Chen and Moritz Voss "On Parametric Optimal Execution and Machine Learning Surrogates" is at https://lnkd.in/gjX5E9D7We apply neural networks to learn optimal execution strategies with transient nonlinear price impact, inventory penalty and stochastic resilience. Given that there are many parameters, we embed them as inputs into the NN surrogate, so that the modeler gets not just one solution, but all of them across a range of model configurations (which enables insights into solution sensitivity, as well as robust approaches -- stay tuned for another paper on that soon).We also put in the effort to make an accompanying Jupyter Notebook (https://lnkd.in/g6N8hCZs) -- not only to illustrate the algorithm and make it reproducible, but also because we believe this is a perfect case study for anyone getting started with using NNs for stochastic control.Finally, a word of thanks to Charles-Albert Lehalle whose very nice lecture at WCMF10 back in early 2021 was one of the motivators for our project. #MLfinance #hft
2022/05/13
25
1
3
Marco Tavora Ph.D.
Open
https://www.linkedin.com/feed/update/urn:li:activity:6930964844867112960
Principia Technologies | Founder - Entrepreneur
https://www.linkedin.com/in/marco-tavora/
The following quote by Einstein always inspires me, no matter how often I see it.In a 1929 interview, he was asked:"How did he get his ideas?"His answer was:"I'm enough of an artist to draw freely on my imagination. Imagination is more important than knowledge. Knowledge is limited. Imagination encircles the world."#creativity #innovation #einstein
2022/05/13
35
2
2
https://media-exp1.licdn.com/dms/image/C4D22AQEQ8vwUdZbyOA/feedshare-shrink_800/0/1652470789407?e=1655337600&v=beta&t=3aCa0hf7m1MadJk1Kvuq8cE_AmZuDSbcFQpPD9JQdcA
Markus Leippold
Open
https://www.linkedin.com/feed/update/urn:li:activity:6930945539756711936
University of Zurich - Department of Banking and Finance | Professor, Chair in Financial Engineering
https://www.linkedin.com/in/markus-leippold-578bb95/
According to the the Research Excellence Framework (REF), a research impact evaluation of British higher education institutions, Imperial has a greater proportion of world-leading research than any other UK university. https://lnkd.in/eWPWCBmG #research
2022/05/13
11
1
2
Matthew Dixon
Open
https://www.linkedin.com/feed/update/urn:li:activity:6930361958030270464
CFX Labs Inc. | Co-Founder
https://www.linkedin.com/in/mfrdixon/
For me personally, one of the holy grails is to tie machine learning and finance to climate change - that is using ML in finance to tackle some of the most challenging problems of our time. In addition to ESG investing, providing financial protection against climate risk is another key ingredient - we need to protect society while figuring out our sustainability transition plan. Natural disasters could otherwise wipe out budgets set aside for R&D, e.g. Hurricane Katrina devastated New Orleans. That, perversely, over time gives fossil fuel companies the upper hand.Destruction -> poverty-> lack of education-> poor personal sustainability hygiene and lack of local opposition to fossil fuel companies (fossil fuel jobs>perceived environmental damage)While not a climate solution, there is a growing contingent claim product area for managing climate change and natural disaster risk and one area where richer datasets (e.g. satellite imagery) and ML can ultimately make a significant difference in ensuring fair pricing and more robust uncertainty quantification. Our latest technical paper on embedding clustering into a unified Hierarchical Bayesian modeling framework for catastrophe and interest rate risk premia adjusted CAT bond pricing is available on arXiv: https://lnkd.in/dQivqGXW This is joint work with co-authors Chatterjee and Domfeh. #climaterisk #machinelearning #environmentalfinance #bayesianstatistics #insuretech #quantitativefinance #weatherforecasting Peter Adriaens Dixon Domfeh Morton Lane Runhuan Feng, PhD, FSA, CERA Todd Ringler Swami Sethuraman Joydeep Lahiri Maura Feddersen Larry Eisenberg Dr. Sebastian Rath Richard Matsui Lawrence Habahbeh
2022/05/12
92
7
5
Saeed Amen
Open
https://www.linkedin.com/feed/update/urn:li:activity:6930458226031271936
Cuemacro | Founder
https://www.linkedin.com/in/saeedamen/
“In this period of dramatic structural change, I do not care about technology. Tech is the least important part of the next 20 years.” Paul Donovan, Chief Economist at UBS Global Wealth Management, said at Calcalist’s Meet &Tech event. “The economic transformation, the economic change, does not come from technology, but how we use it.”
2022/05/12
82
4
4
https://media-exp1.licdn.com/dms/image/sync/C4D27AQH0zKce0poylQ/articleshare-shrink_1280_800/0/1652364084773?e=1652475600&v=beta&t=i_qV_twpRir0OtoJzrRjQGsgpcdIUl_0TWHfzRNPGhc
Matthew Dixon
Open
https://www.linkedin.com/feed/update/urn:li:activity:6930591654051557376
CFX Labs Inc. | Co-Founder
https://www.linkedin.com/in/mfrdixon/
BlackRock and Citadel allegedly caused one of the largest #crypto crash of the last years. The two investors allegedly borrowed 100k #BTC from Gemini, swapped 25K of #BTC into #TerraUSD (#UST), called Terra Foundation suggesting they’d want to sell a large chunk of BTC for UST, and hinted to buy them at discount to avoid moving the market. https://lnkd.in/dAi_yJm2Terra ($Luna) agreed to buy BTC for UST at a discount, lowering its own UST liquidity and reserves significantly. After which, BlackRock and Citadel dumped all UST and #BTC causing massive slippage and triggering a cascade of forced selling in both assets. At the same time, this also triggered a cascade of withdrawals from another large crypto, $AnchorProtocol, holding large quantities of $Luna. The withdrawals were more than Anchor could repay, triggering a further sell-off in $Luna, thus further breaking the $1 peg of UST/USD.#BTC currently trading below $27,000 (down 32.2% in 1 week) can now be bought by BlackRock and Citadel much cheaply, they can repay their loan, and pocket billions in difference. Meanwhile, over $300bn of long positions in stable and altcoins were wiped out, and $Luna tumbled from $86 to $0.16 in less than a week. While BlackRock and Citadel denied the claims, only the actions of very large institutional investors could trigger such a tumble by two of the largest #cryptocurrencies. This was pure market manipulation. It is surprising that the largest investors in the world and loud advocate for ESG and Ethical investment is allegedly behind such actions. At which point trading and investing becomes market manipulation? Is #crypto investing supposed to align to #ESG standards, or is it still considered no man's land due to lack of regulations? #bitcoin #trading #investment #crypto #responsibleinvestment #blockchain #cryptocurrencies #altcoins
2022/05/12
1358
318
3
https://media-exp1.licdn.com/dms/image/C4D22AQHclIyCFIo5Lw/feedshare-shrink_2048_1536/0/1652342960748?e=1655337600&v=beta&t=GhpqCUDsUmPMCMsasF2Hc52heEqlVYMQNe7KDuZt-IA
Charles-Albert Lehalle
Open
https://www.linkedin.com/feed/update/urn:li:activity:6930552429025325056
Abu Dhabi Investment Authority (ADIA) | Quantitative R&D Lead
https://www.linkedin.com/in/lehalle/
Best explanation of how algorithmic stablecoins work I've seen yet as well as a logical explanation of possible outcomes. "Here is how an algorithmic stablecoin works:1. You wake up one morning and invent two crypto tokens.2. One of them is the stablecoin, which I will call “Terra,” for reasons that will become apparent.3. The other one is not the stablecoin. I will call it “Luna.”4. To be clear, they are both just things you made up, just numbers on a ledger. (Probably the ledger is maintained on a decentralized blockchain, though in theory you could do this on your computer in Excel.) 5. You try to find people to buy them.6. Luna will trade at some price determined by supply and demand. If you make it up on your computer and keep the list in Excel and smirk when you tell people about this, that price will be zero, and none of this will work.7. But if you do a good job of marketing Luna, that price will not be zero. If the price is not zero then you’re in business.8. You promise that people can always exchange one Terra for $1 worth of Luna. If Luna trades at $0.10, then one Terra will get you 10 Luna. If Luna trades at $20, then one Terra will get you 0.05 Luna. Doesn’t matter. The price of Luna is arbitrary, but one Terra always gets you $1 worth of Luna.... 9. You set up an automated smart contract — the “algorithm” in “algorithmic stablecoin” — to let people exchange their Terras for Lunas and Lunas for Terras.10. Terra should trade at $1. If it trades above $1, people — arbitrageurs — can buy $1 worth of Luna for $1 and exchange them for one Terra worth more than a dollar, for an instant profit. If it trades below $1, people can buy one Terra for less than a dollar and exchange it for $1 worth of Luna, for an instant profit. These arbitrage trades push the price of Terra back to $1 if it ever goes higher or lower.11. The price of Luna will fluctuate. Over time,
2022/05/12
114
25
3
https://media-exp1.licdn.com/dms/image/sync/C4D27AQGAjcNFNCqVPw/articleshare-shrink_800/0/1652359849508?e=1652994000&v=beta&t=sDwb9elLlGNForezeEAnGLroDYtAoMUm0vKnjK0WZ-k
Saeed Amen
Open
https://www.linkedin.com/feed/update/urn:li:activity:6930577349834895360
Cuemacro | Founder
https://www.linkedin.com/in/saeedamen/
Turnleaf Analytics is hiring data scientists/engineers and interns! If you'd be interested in working on forecasting of EM inflation, using cutting edge techniques, with Alexander Denev and me, check the link below! #datascience #careers
2022/05/12
28
2
3
Wesley Gray, Ph.D.
Open
https://www.linkedin.com/feed/update/urn:li:activity:6930520043319349248
Alpha Architect, LLC | CEO
https://www.linkedin.com/in/alphaarchitect/
With our mission of educating people to become better #investors, we are revolutionizing asset management.We'd love to hear from anyone interested in joining our team to help us out with #compliance as our Deputy CCO.Please see link in comments field for job description.
2022/05/12
18
1
3
https://media-exp1.licdn.com/dms/image/C4D22AQFrJsKqhvQ_yQ/feedshare-shrink_800/0/1652364739779?e=1655337600&v=beta&t=KGV28R5HoXlJggnuDNU0KIRbhslgQ-9vrTEhms5c9aY
Prof. Alexander Lipton
Open
https://www.linkedin.com/feed/update/urn:li:activity:6930465927444467712
Sila | Co-Founder, Chief Information Officer
https://www.linkedin.com/in/prof-alexander-lipton-aa2256bb/
What can we do with NFTs besides trade? Did you ever wonder if Zero-Value NFTs make sense? Together with Thomas Hardjono we propose a computation model - called the Asset Proxy NFT paradigm - that guarantees consistency between on-chain and off-chain data. The second contribution of this paper is the notion of the NFT design patterns which recognizes that a coherent framework for dealing with hybrid (on-chain & off-chain) assets is required. PDF of the full paper available here https://lnkd.in/dEktPVBB#hybridNFTs #hybridAssets #NFT #blockchain #Web3 Photo Credit (https://lnkd.in/dDyhYD4u)2022/05/12
32
14
2
https://media-exp1.licdn.com/dms/image/C4D22AQF4tNln-lUoXw/feedshare-shrink_2048_1536/0/1652350370291?e=1655337600&v=beta&t=B_gwnXwn-DxLLQpWa1owrPDS3vykw9BjNwDbndG8KXg
Peter Cotton, PhD
Open
https://www.linkedin.com/feed/update/urn:li:activity:6930191486651179008
Intech Investment Management LLC | Senior Vice President. Chief Data Scientist.
https://www.linkedin.com/in/petercotton/
Coinbase said in its earnings report Tuesday that it holds $256 billion in both fiat currencies and cryptocurrencies on behalf of its customers. Yet the exchange noted that in the event it ever declared bankruptcy, “the crypto assets we hold in custody on behalf of our customers could be subject to bankruptcy proceedings.” Coinbase users would become “general unsecured creditors,” meaning they have no right to claim any specific property from the exchange in proceedings. Their funds would become inaccessible.Coinbase is down 70% YTD.FUD?
2022/05/11
53
52
5
https://media-exp1.licdn.com/dms/image/sync/C5627AQFqMaDeShX3GA/articleshare-shrink_800/0/1652299558709?e=1652389200&v=beta&t=3q6s4ZFGhEPvBeu9VR14dZuC-0U5DNSt4kEv_FcdIB8
Álvaro Cartea
Open
https://www.linkedin.com/feed/update/urn:li:activity:6930141738711855104
University of Oxford | Professor
https://www.linkedin.com/in/alvaro-cartea/
New paper "Algorithmic Collusion in Electronic Markets: The Impact of Tick Size". This is joint work with Patrick Chang (Oxford-Man Institute) and Jose Penalva (Universidad Carlos III de Madrid & Associate member of the Oxford-Man Institute). All comments welcome. #algorithmictrading #machinelearning #machinelearningalgorithms #microstructure #regulation #reinforcementlearning #gametheory #quantitativefinance #artificialintelligence
2022/05/11
64
0
4
Spyros Makridakis
Open
https://www.linkedin.com/feed/update/urn:li:activity:6930141234766221312
University of Nicosia | Professor
https://www.linkedin.com/in/spyros-makridakis-b2ba5a52/
The M6 financial forecasting competition is going strong, now in its third month with more than 200 participating teams. The Figure below shows the graph of the IR (IR= returns/sd, see guidelines) adjusted by the value of benchmark is subtracted from IR. It is interesting to see the learning process as the percentage of those beating the benchmark is improving continuously over time (shown by the increasing number of teams above the zero line and their higher IRs).Evangelos Spiliotis
2022/05/11
55
5
3
https://media-exp1.licdn.com/dms/image/C4E22AQHE036iuGD-lw/feedshare-shrink_800/0/1652274424817?e=1655337600&v=beta&t=mJceTkWSIqtGM0DhO_kMG78_UNlGgYqRkm58oa1c8mc
Igor Halperin
Open
https://www.linkedin.com/feed/update/urn:li:activity:6930131173276889088
Fidelity Investments | AI Asset Management
https://www.linkedin.com/in/igor-halperin-092175a/
Delighted to share Barron's article on AI and wealth management featuring my paper with Matthew Dixon, which shows how reinforcement learning (RL) and inverse reinforcement learning (IRL) can be combined for wealth management for retail investors.A related paper that proposes combining RL and IRL for optimization of asset allocation by professional portfolio managers is available here: https://lnkd.in/dQqXz-Pt #investing #wealthmanagement #quantitativefinance #reinforcementlearning #inversereinforcementlearning
2022/05/11
52
0
3
Matthew Dixon
Open
https://www.linkedin.com/feed/update/urn:li:activity:6930228390310424576
CFX Labs Inc. | Co-Founder
https://www.linkedin.com/in/mfrdixon/
Delighted to see our Illinois Tech spin-out, CoinFX, listed as one of the top 20 innovators in web3 and crypto by Forbes. CoinFX is a registered money services business that facilitates stablecoin minting in 130 local currencies backed by USDC and USDP: https://lnkd.in/gfmUr4ve A big thanks to Nik Rokop and Maryam Saleh at Illinois Tech, Paul Hsu and Allister Lam at Decasonic, and our investor network listed at coinfx.net. Built on the solana blockchain and designed with the discipline of rigorous mathematical modeling and stress testing using risk management best practices from investment banking. #stablecoins #riskmanagement #fintech #fintechstartup #foreignexchange #crossborderpayments #blockchain #solana
2022/05/11
21
4
3
https://media-exp1.licdn.com/dms/image/sync/C5627AQHC7DT7Xz1egA/articleshare-shrink_800/0/1652299641549?e=1652389200&v=beta&t=QjnuZ9p7yWhgu3JcVKIHjKw5NpQpoTuZIzzzB4QX22k
Load 50 more
TOP