SYSTEM MEDICINE

   The exponential grow of biological data and companion models suited to make progress

on medicine need comprehensive logical methods to transform the medical practice

on a rigorous repeatable pattern.

    The operational protocols serving clinical medicine, even in high standards hospitals, are conceived on general approaches, the treatment, reducing to same well know schema, often missing personalizing elements.

The reality is the clinician has not a lot of choice or freedom to innovate therapies, in the con-

text where strict legislation is applied to overcome mistakes.

Translational Medicine approach is suited, but this is seen quite sparse in the hospitals and usually the methods are empirical. In diagnostic modeling, one has to interpret not only the result of

the model, but the way the result has been obtained.

The robust MEDICOMPILLS technology is needed, at least for three reasons:

1. to solve correctly medical issues

2. to give easy to derive value of all molecular biology, bioinformatics, computational

biology, biomedecine knowledge acquired over the last 30 years

3. as a tool which could be entitled as standard explanatory reference for legal

agreement.

Despite the LLM tools, where the risk of generating cascade of errorsis high, this

new compiler approach has a traceability feature and an a priori potential to achieve

an Exact Medicine.

PUblications

1.   A. Climescu-Haulica, MEDICOMPILLS: a Meta Architecture to Build Robust Therapeutic

and Diagnostic Systems, 2025

2.   A. Climescu-Haulica, MEDICOMPILLS Architecture: a Mathematical Precise Tool to

Reduce the Risk of Diagnosis Errors on Precise Medicine, XVII International Conference

 in  Precision Medicine, Montreal, Canada, June , 2023



3. A. Climescu-Haulica, Machine learning algorithms on multi-layer architecture to process

hidden information for systems medicine applications, AMLD January 2019, DOI:

10.13140/RG.2.2.23604.09606

Machine Learning

Mathematical Machine Learning is our proposal for high tech  companies looking for roboustness and precision.

                                 Our protocols are rigorously based on:

1.Proofs

2.The diversity of methods

3.Smart testing examples

Originality & Optimality

PUBLICATIONS

 1. A. Climescu-Haulica  Machine Learning to build Connectomics Architecture from

Stochastic Differential Equations with Jump Noise January 2025

2. A. Climescu-Haulica How to choose the number of clusters : the Cramer multiplicity solution in ”Advances in

Data Analysis ” Studies in Classification, Data Analysis, and Knowledge Organization, R.Decker and H-J. Lenz

(editors) Springer (2007) http://www.springerlink.com/content/71471260833634

3. G. Alexe, G. Bhanot, A. Climescu-Haulica A cross entropy algorithm for classification with delta-patterns Discrete Mathematics and Theoretical Computer Science (2006) vol. AG 399-402 https://hal.archives-ouvertes.fr/hal-01184688v1

4. A. Climescu-Haulica Learning in Spike Neuronal Models 13th INFORMS Applied Probability Conference Ottawa (2005) https://informs.emeetingsonline.com/emeetings/informs/144/paper/11963.

5. Michelle Quirk, A. Climescu-Haulica, Kevin Saeger Network control by normalized cut fuzzy clustering strategy for critical infrastructures Proceedings of Third International Conference on Computing, Communication and Control Technologies (2005) Austin

6. A. Climescu-Haulica Large Deviation Analysis of Space-Time Trellis Codes in Trends in Mathematics, ”Mathe-

matics and Computer Science” Algorithms, Trees, Combinatorics and Probabilities, M. Drmota, P. Flajolet, D.

Gardy, B. Gittenberger (editors) Birkhauser (2004)

http://www.springer.com/east/home/birkhauser/mathematics

SIGNAL PROCESSING

Mathematics for Signal Processing evolved in the last 30 years but their practical implementation is still sparse. The development of Signal Processing  is lately feed more by Machine Learning techniques than from new spectral methods. Our research covers the gap needed to make innovation a constant improvement in communication area.

PUBLICATIONS:

1.A.F. Gualtierotti , A.Climescu-Haulica , M.D. Pal Likelihood ratio detection of signals in reverberation noise

Proceedings of IEEE International Conference in Acoustics, Speech and Signal Processing (2002) Orlando pages 1589-159 http://www.cmsworldwide.com/ICASSP2002/

2. A. Climescu-Haulica, A.F. Gualtierotti The Likelihood ratio for the detection of a random signal of an unknown

law, embedded in altered and weighted Wiener and Poisson noises Proceedings of IEEE International Symposium on Information Theory (2002) Lausanne http://www.isit02.epfl.ch

3. C.R. Baker, A. Climescu-Haulica, A. F. Gualtierotti Detection of random signals in Gaussian and NonGaussian

dependent noise Proceedings of Fifth IMA International Conference on Mathematics in Signal Processing (2000) Warwick pages 345-349

ComputaTional Biology

Next generation computational biology is needed to assistnthe emergent field of molecular system medicine. The development speed of omics tools fuels the increase of multi-omics knowledge databases,moving the medicine towards molecular grounds. The question is the

integration of omics with physiology, for which new conceptual methods are in demand. Our work comes from the following observation: the bio-molecular relationships are extremely complex while the investigating methods, such as bio-networks and features extraction from bio-data sets,

are too restrictive to model living organism complexity.

Important dynamical information is lost between. The knowledge acquired until now is

appealing to infer, using advanced mathematical objects, more biological insights, so better medical logic. With this vision we introduce a pioneering approach which address a central concept: the biological connectivity.

Indeed, the network model, an engineering object, is too schematic to de-

scribe connection in living system sense. Our findings may have a key role on for solving open problems in computational biology.

Publications

1.A. Climescu-Haulica Biocompilation Assembles Languages to Express Biological

.Connectivity, ISMB JUly 2024, Montreal

2. A. Climescu-Haulica, DataCompil – a Multimodal Integration to Leverage Biological

Connectivity, RECOMB, Boston, May 2024

3. R. Turliuc, B. Grecu, A. Climescu-Haulica Dynamic classification of the human gene damage index: a physiological approach 20th Annual International Conference on Research in Computational Molecular Biology (2016) Santa Monica, April 17-21

4. A. Climescu-Haulica Integrative biomarkers for omics data BioIT World Conference (2009) World Trade Center Boston, April 27-2

5. A. Climescu-Haulica Accurate biomarkers selection from large data set Biomarkers Europe (2008) Vienna, 10-11 November

6.. A. Climescu-Haulica, M.Quirk Nonlinear stochastic differential equations method to infer gene regulatory network architecture 6th Workshop on Statistical Methods for Post-Genomic Data (2008) Rennes, 31 January-1 February

7. A. Climescu-Haulica, M.Quirk Nonlinear stochastic differential equations method for reverse engineering of gene regulatory networks in Computational Methodologies in Gene Regulatory Networks S. Das, D. Caragea, W. H.Hsu, S. M. Welch (editors) Information Science Publishing (2008)

8. A. Climescu-Haulica, M. Quirk A Fuzzy Games Approach of RNA-based gene regulation RECOMB(2007) San

Francisco http://www.recomb2007.com/posters/climescuRECOMB2007.pdf

9. A. Climescu-Haulica, M.Quirk A nonlinear dynamical model to infer transcriptional regulatory networks from time dependent expression data, 4th Annual Recomb Satellite on Regulatory Genomics, MIT Broad Institute (2007) 11-13 October http://compbio.mit.edu/recombsat/2007/proceedings.pdf

10.  A. Climescu-Haulica, M. Quirk A stochastic differential equation model for transcriptional regulatory networks.BMC Bioinformatics 8, S4 (2007) https://doi.org/10.1186/1471-2105-8-S5-S4

11. G. Alexe, G. Bhanot, A. Climescu-Haulica Accurate classification of cancer phenotypes via an entropy based Monte Carlo method 15th Annual International Conference on Intelligent Systems for Molecular Biology (2007) Vienna, July 21-22 http://www.iscb.org/cmsaddon/conferences/ismb/poster˙list

12. A. Climescu-Haulica, M. Quirk A stochastic differential equation model of trancriptional regulatory network RECOMB (2006) Venice

NEUROSCIENCE

Our research in neuroscience in driven by three factors:

1. to reveal mathematically the importance of the neuro-endocrine system for the human health 

2. to investigate the increasing neuro-data with right computational methods

3. to model brain typologies as essential tool for personalized medicine, integrating neuroscience in system medicine

Publications:

1.A. Climescu-Haulica, The Brain's Essential Role in Mediating Immune Responses: HPA Axis to Leverage Signals with a Systemic Approach. BRAIN. Broad Research in Artificial Intelligence and Neuroscience, [S.l.], v. 15, n. 4, p. 375-386, dec. 2024. ISSN 2067-3957. Available at: <https://brain.edusoft.ro/index.php/brain/article/view/1647/1955>.

2. C. La Rota, D. Vuza, A. Climescu-Haulica, Hierarchical Model of the

Hypothalamo-Adreno-Pituitary system  https://doi.org/10.1101/837724,

Cold Spring Harbor laboratory, bioXxiv, 46 pages, October 2019

3. A. Climescu-Haulica A Stochastic Calculus Approach of Learning in Spike Models Workshop on Spike Time Dependent Plasticity (2004) Monte Verit`a http://icwww.epfl.ch/ gerstner/STDP/index.ht

STOCHASTICS   &  APPLICATIONS

Designed for Signal Processing, Machine Learning, Neuroscience or Computational Biology, our expertise in stochastic calculus brings key elements to detect signal from noise, whatever form they may have.

This is a very interesting mathematical journey, adding new advances in fundamental mathematics as well as in applications, seen always in their realistic, society serving achievement.

Publications:

1.A. Climescu-Haulica Voiculescu’s free entropy and spectral analysis of random graphs The Eighth Congress of Romanian Mathematicians (2015) Iasi, June 29 - July 1

2. A. Climescu-Haulica Filtrage stochastique non lin´eaire par la th´eorie de repr´esentation des martingales Proceedings XXXVI-`emes Journ´ees de Statistiques (2004) Montpellier       

www.agromontpellier. fr/sfds/CD/textes.html

3. A. Climescu-Haulica and A. F. Gualtierotti Likelihood ratio detection of random signals : the case of causally

altered and weighted Wiener and Poisson International Journal of Pure and Applied Mathematics (2002) vol. 2,

pages 155-217

4.L. Caramellino, A. Climescu-Haulica, B. Pacchiarotti Diffusion approximation for random walks on nilpotent Lie groups Statistics and Probability Letters (1999) vol. 41, pages 363-377

Connectivity without network: ask about our  recent innovation

Optimal architecture taylored to client application